(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Method for minimizing technetium-99 migration from a geologic repository"

METHOD FOR MINIMIZING TECHNETIUM-99 MIGRATION 
FROM A GEOLOGIC REPOSITORY 









VIRLYNDA DEMARTINO 



A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL 

OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT 

OF THE REQUIREMENTS FOR THE DEGREE OF 

DOCTOR OF PHILOSOPHY 

UNIVERSITY OF FLORIDA 

2000 









Copyright 2 000 
VirLynda DeMartino 






TO THE TWO PEOPLE WITH THE GREATEST INFLUENCE ON MY LIFE: 



MY DEAREST MOTHER 
WITH ETERNAL LOVE AND ADMIRATION 



JAMES HORAK 
"ALWAYS ACHIEVE YOUR MAXIMUM POTENTIAL." 



ACKNOWLEDGMENTS 



I first and foremost thank God, for it is only through 
Him that all things are made possible. 

I thank the U.S. Department of Energy for their 
financial support administered by AWU-NW, NORCUS, and ORAU 
Programs. I thank my committee cochairmen: S. Anghaie 
(University of Florida) and W.J. Gray (Pacific Northwest 
National Laboratory), and my committee members: 
G. Schoessow, J. Tulenko, and R. Hanrahan. I thank 
L. Greenwood (PNNL) and C. Wilson (Westinghouse 
Corporation) . I thank W.H. Ellis, my original cochairman, 
who retired before I received this degree. 

Extra special recognition is given to those listed 
below. These are my loved ones who were always there for me 
throughout this endeavor. Their love and support kept me 
going, and they share my joy in receiving this degree. 

I thank my husband, who has always given me emotional 
support. He helped me to realize and utilize my 
internal strength to face all life's trials and 
tribulations, never looking for recognition or thanks, 

iv 



only what was best for me. His confidence in me has 
strengthened me and given me the confidence to always 
achieve my maximum potential in all of life's 
endeavors. 

I thank my daughter, who has had to be without her 
Mommy many hours while I worked on this degree. Her 
bright smile and loving ways have definitely been the 
highlight of many gloomy hours! 

I thank my mother, who has given undying support 
throughout my entire life. Her love, encouragement, 
and reassurance have given me the ambition to keep my 
dreams and goals alive. Though many miles away, she 
was always there enduring my hardest times, as well as 
triumphing in the more glorious times; always keeping 
me striving for better and higher goals. 

I thank my sister, who gave me immeasurable amounts of 
encouragement and love throughout this adventure. 

I thank James A. Horak whose fatherly love led me to 
choose this discipline by the tender age of nine. His 
words of "always strive to achieve your maximal 
potential" kept me striving to complete the Ph.D. 



I thank my family, who were always ready to offer their 
advice when the going got tough. I especially thank my 
sister for all her prayers and constant visits to the 
Grotto. 

I thank my numerous friends not individually mentioned, 
who, nonetheless, have been a great support over the 
years. I thank them especially for their prayers. 



vi 



TABLE OF CONTENTS 

ACKNOWLEDGMENTS iv 

LIST OF FIGURES X 

LIST OF TABLES xiii 

ABSTRACT xiv 

'"■'.*\. ; '■■':■>: 

CHAPTERS ' ' , .'. 

1 INTRODUCTION 1 

1.1 Technetium Background Information 1 

1.2 Computational-Analysis Approach 6 

1.3 Experimental Approach 10 

1.4 Laboratory Closure 11 

1.5 Development of This Dissertation 12 

2 COMPUTATION ANALYSIS AND RESULTS 13 

2.1 Modeling the System 13 

2.2 The EQ3/6 Computer Codes 20 

2.3 EQ3/6 Simulations 22 

2.3.1 Preliminary EQ3/6 Simulations .... 23 

2.3.2 Interpreting the EQ3/6 Figures .... 30 

2.3.3 Simulations Modeling the Experiments . 35 

2.4 EQ3/6 and the Experimental Results 41 

3 EXPERIMENTAL PREPARATION AND DESCRIPTIONS 45 

3.1 Technetium Solutions 45 

3.2 Technetium-Detection System 46 

3.3 Material Acquisition 47 

3.4 Systems Used and Experiment Descriptions . . 48 

3.4.1 FT-4C Flow-Through System 54 

3.4.1.1 Equipment description 54 

3.4.1.2 Equipment readiness experiment 57 

3.4.1.3 Experiments' descriptions ... 59 

3.4.2 Bench-Top Flow-Through System .... 61 

3.4.2.1 Equipment description 62 

3.4.2.2 Bench-top experiments 62 

3.4.3 Beaker System 63 

3.4.3.1 Equipment description 64 

3.4.3.2 Experiment description .... 64 

3.4.4 Sealed Vessel System 66 

3.4.4.1 Equipment description 66 

3.4.4.2 Readiness experiments 68 

vii 



3.4.4.3 Experiment description .... 69 

4 RESULTS AND DISCUSSION 72 

4.1 Overview 72 

4.2 Blank-run Experiment 75 

4.3 Iron Powder Experiments 81 

4.3.1 Bench-Top Iron Powder Experiments . . 81 

4.3.2 Sealed Vessel Experiments 93 

4.3.2.1 Sealed Vessel Experiment at 

70 °C 94 

4.3.2.2 Sealed Vessel Experiment at 

24°C 105 

. 4.3.3 FT-4C Iron Powder Experiments .... 113 

4.3.3.1 High-flow Iron Powder 

Experiment 114 

4.3.3.2 Low-flow Iron Powder 

Experiment 117 

4.3.3.3 Iron Powder Experiment at 85°C 118 
4.3.4 Iron Powder Summary and Conclusions . 121 

4.4 Copper Experiments 12 5 

4.4.1 Bench-Top Copper Powder Experiment . . 126 

4.4.2 FT-4C Copper Shot Experiments .... 136 

4.4.2.1 DIW Copper Shot Experiment . . 136 

4.4.2.2 Copper Shot Experiment at 24 °C 139 

4.4.2.3 Copper Shot Experiment at 85°C 14 3 

4.4.3 Copper Summary and Conclusions .... 144 

4.5 Stannous Chloride Experiments 147 

4.5.1 Beaker Stannous Chloride Experiment . 148 

4.5.2 Bench-Top Stannous Chloride 

Experiment 152 

4.5.3 Stannous Chloride Summary and 
Conclusions 156 

4.6 FT-4C Tin Shot Experiments 159 

4.6.1 Tin Shot Experiment at 24°C 160 

4.6.2 Tin Shot Experiment at 85°C 160 

4.6.3 Tin Summary and Conclusions 164 

5 SUMMARY AND CONCLUSIONS 166 

6 RECOMMENDATIONS 172 

7 EPILOGUE 177 

GLOSSARY 179 

APPENDICES 

A THE EQ3/6 COMPUTER CODES 185 

B BETA DETECTION SYSTEM 196 

C GAMMA DETECTION SYSTEMS 200 



Vlll 



D CHEMICAL ASSAY DATA ' 2 09 

REFERENCES 214 

BIOGRAPHICAL SKETCH 221 






4 i .i''* 



ix 



LIST OF FIGURES 



Figure page 



1 EQ3/6 Copper Simulation Solution Environment ... 25 

2 EQ3/6 Copper Simulation Precipitated Technetium 
Compounds 26 



« ^ 



3 EQ3/6 Copper Simulation Solution Environment and 
Precipitated Technetium Compounds 27 

4 EQ3/6 Stannous Chloride Simulation Solution Environment 

28 

5 EQ3/6 Stannous Chloride Simulation Precipitated 
Technetium Compounds 29 

6 Fluid-centered Flow-through Open System 31 

7 EQ3/6 Simulation Technetium Concentrations and Product 
Minerals for Iron Reactant 3 3 

8 EQ3/6 Iron Simulations 36 

9 EQ3/6 Closed System Simulation Technetium 
Concentration 39 

10 EQ3/6 Closed System Simulation Solution Environment 40 

11 Experimental Systems 49 

12 Translational Column Assembly Cutaway View .... 51 

13 Detailed Diagram of the FT-4C Flow-Through System. 56 

14 FT-4C Readiness Experiment Using Uranium 58 

15 Sealed Vessel System Eh Instrumentation Calibration 
Using Quinnhydrone 70 

16 FT-4C Blank-run Experiment 78 



17 Bench-top "Tc Iron Powder Experiment 82 

18 Bench-top "^^"^c Iron Powder Experiment Solution 
Concentrations 84 

19 Bench-top '^mr^^ iron Powder Experiment Reaction 

Column 86 

20 Comparison of the Bench-top ^^"^c Iron Powder Experiment 
and EQ3/6 Simulations 90 

21 Sealed Vessel Experiment at 70°C and the EQ3/6 
Simulation Technetium Concentrations and Predicted 
Precipitation of Technetium Compounds 95 

22 Sealed Vessel Experiment at 70°C and the EQ3/6 
Simulation Solution pH 96 

23 Sealed Vessel Experiment at 70°C and the EQ3/6 
Simulation Solution Environment 97 

24 Sealed Vessel Experiment at 70°C and the EQ3/6 
Simulation Solution pH and Eh 98 

25 Sealed Vessel Experiment at 24 °C and the EQ3/6 
Simulation Technetium Concentrations and Predicted 
Precipitation of Technetium Compounds 107 

26 Sealed Vessel Iron Powder Experiment at 24 °C and the 
EQ3/6 Simulation Solution pH 108 

27 Sealed Vessel Iron Powder Experiment at 24 °C and the 
EQ3/6 Simulation Solution Environment 109 

28 Sealed Vessel Iron Powder Experiment at 24 °C and the 
EQ3/6 Simulation Solution pH and Eh 110 

29 FT-4C High-flow Iron Powder Experiment 115 

30 FT-4C Low-flow Iron Powder Experiment 119 

31 FT-4C Iron Powder Experiment at 85°C 120 

32 Bench-top '^mrp^ Copper Powder Experiment and the FT-4C 
Copper Shot Experiments Using J-13 128 

33 Comparison of the Bench-Top Copper Powder Experiment 
and the EQ3/6 Simulation 131 

34 Bench-top Copper Powder Experiment and the EQ3/6 
Simulation Using Aqueous Copper 134 

xi 



35 FT-4C DIW Copper Shot Experiment 137 

36 FT-4C Copper Shot Experiments at 24 °C and 85°C . . 140 

37 EQ3/6 Simulation Technetium Concentrations and Product 
Minerals for Copper Reactant 141 

38 Beaker and Bench-top Stannous Chloride Experiments 149 

39 Comparison of the Beaker Stannous Chloride Experiment 
and the EQ3/6 Simulation 151 

40 Comparison of the Bench-Top Stannous Chloride 
Experiment and EQ3/6 Simulation 155 

41 FT-4C Tin Shot Experiments at 24 °C and 85°C . . . 161 

42 Comparison of the Tin Shot Experiments and the EQ3/6 
Simulation 162 



Xll 



LIST OF TABLES 

Table page 

1 Pressurized Water Reactor Spent-Fuel Radionuclide 

Inventories at 1,000 Years 4 

2 J-13 Well Water Composition 8 

3 Properties of Technetium Isotopes of Interest ... 46 

4 Experiments Conducted 52 

5 Experiment Location 73 

6 Experimental Results 74 

7 Chronology of FT-4C Experiments 76 



Xlll 



Abstract of Dissertation Presented to the Graduate School 

of the University of Florida in Partial Fulfillment of the 

Requirements for the Degree of Doctor of Philosophy 

METHOD FOR MINIMIZING TECHNETIUM-99 MIGRATION 
FROM A GEOLOGIC REPOSITORY 

By 

VirLynda DeMartino 

December 2000 

Chairman: Samim Anghaie 

Cochairman: Walter J. Gray 

Major Department: Nuclear and Radiological Engineering 

■ A method for minimizing technetium-99 ("tc) migration 

in groundwater by using reducing agents was investigated. 

This information could be used to tailor the spent nuclear 

fuel waste packages or near-field environment at a geologic 

repository to prevent ''Tc migration should water 

infiltration occur. The proposed geologic repository at 

Yucca Mountain will likely remain dry for several hundred 

years because (1) it will be located in the unsaturated zone 

above the water table, and (2) it will have significant 

radioactive decay heat. As time passes, the decay heat 

generated will lessen. The repository will then cool below 

the boiling point of water, and then water could contact 

spent fuel in any breached waste package. At that time, the 



XIV 



soluble radionuclides, such as ''tc, may possibly dissolve 
and migrate from the repository into the environment. 

This research also could be applied to technetium-soil 
decontamination by adding the reducing agent to the soil and 
disposing of the soil as hazardous mixed waste. 

A computational analysis (EQ3/6 software package) and 
an experimental approach were used during this 
investigation. Both approaches investigated the effect of 
reducing agents on technetium (VII) solutions. 
Theoretically, the ''tc would precipitate as insoluble, 
lower valent oxides or technetium metal because of the 
reducing environment, and its migration would be minimized. 

Copper, iron, stannous chloride, and tin were used as 
reducing agents in both approaches. Experimentally, iron 
reduced the technetium concentration to below detectable 
limits; copper, tin, and stannous chloride only temporarily 
reduced the technetium concentration. 

The experimental results were consistent with the 
computer simulations for iron and one of the stannous 
chloride experiments. The other stannous chloride 
experiment temporarily reduced the technetium concentration, 
but not to the values indicated by the codes. 
Experimentally, copper and tin developed passivation layers, 
and therefore, the results did not agree with the computer 
simulations. The computer codes should be updated to take 
this into account. Based on the results of these 

XV 



experiments, iron should be considered for reducing the rate 
of technetium migration under oxidizing conditions; stannous 
chloride, copper, and tin should not. 






XVI 



- ' CHAPTER 1 ,'-': '^ 

INTRODUCTION 

1.1 Technetium Background Information 

Perrier and Segre first artificially produced element 
43 in 1937.^'^ They chose the name technetium from the 
Greek word meaning artificial (rexvr^Tos)^ because they 

produced it via the reactions '^Mo42 (d,n) '^""Tc^s and ''^M042 
(d,n) ''''^TC43 using the University of California's 
cyclotron J '^ Since then, technetium isotopes with mass 
numbers from 90 to 111 have been identified, as well as 
isomers with mass numbers 90, 91, 93 - 99, and 102.^ The 
shortest-lived technetium isotope is ^"Tc (0.30 sec 
half-life), and the longest-lived is '^Tc (4,200,000 year 
half-life) .^ However, the three technetium isotopes of 
significant importance today are '^""Tc, '"'"Tc, and ''Tc. 
Technetium-95m and technetium-99m are important in the field 
of nuclear medicine. ''-^ Technetium-99 (213,000-year half- 
life) is important because it is a fission product produced 
in nuclear reactors, with a yield of about 6%.'*'^ Any 
consideration for disposing of spent nuclear fuel must 
account for contributions from technetium-99. 

The U.S. Department of Energy (DOE) is evaluating Yucca 
Mountain, Nevada, as a potential repository site for the 



2 

disposal of spent nuclear fuel, as a waste form.' The 
possibility of releasing radionuclides from the repository 
into the accessible environment must be addressed. The 
scenario by which most of the radionuclides could do so is 
by dissolution and transport via groundwater.^ However, 
the repository will likely remain dry for several hundred 
years because it will be located in the unsaturated zone 
above the water table, and it will be kept hot by 
radioactive decay heat.^ As time passes, the decay heat 
will lessen. Eventually, the repository will cool below the 
boiling point of water, which will allow liquid water to 
contact the spent fuel in any breached waste package. At 
that time, the potential for radionuclide release into the 
environment would exist, and some means of minimizing this 
release is necessary. 

Radionuclides can be divided into three broad 
categories, based on solubility, for modeling their 
potential for release from a repository. ^'^ The two 
considered in this research are low-solubility and high- 
solubility. The third (gaseous radionuclides) was not 
considered in this research.''-^ Solubility constraints will 
limit the rate of groundwater transport of actinides (rare 
earth elements with atomic numbers 89 through 103)' and 
other low-solubility radionuclides away from the 
repository.'''^ Dissolution kinetics of the spent fuel and 
any interactions of the dissolved radionuclides with the 



3 

near-field repository environment will limit the transport 
rate of soluble radionuclides.'''^ 

Technetium-99 falls into the high-solubility 
radionuclide category under the oxidizing conditions 
expected to exist at the potential Yucca Mountain repository 
because it is expected to be present as the pertechnetate 
anion (TCO4") . This radionuclide has a very long half -life 
of 213,000 years. As shown in Table 1, "Tc comprises 0.75% 
of the total 1,000-year activity in the spent fuel, or 13 
Curies per metric ton of heavy metal (MTHM) .^ Other 
investigations have determined that because of the 
solubility of ''Tc, the U.S. Nuclear Regulatory Commission 
(NRC) release limit of one part in 100,000 per year from the 
engineered barrier system could be exceeded if the spent 
fuel is contacted by groundwater . ^°'"'^^'^^'^^'^^'^^ 
Americium and plutonium account for 98.12% of the total 
1,000-year activity, but their release may be limited by low 
solubility.^ 

Some means of minimizing ''tc migration at a geologic 
repository under the scenario of waste-cask breach and 
groundwater infiltration may be necessary to remain within 
the NRC's release limits. The method chosen for 
investigating a means to reduce the rate of ''tc migration 
at a geologic repository was based on the redox 
characteristics of ''Tc applied to standard chemical 
thermodynamic eguations. ^^'^^'^^ 



Table 1 Pressurized Water Reactor Spent- 
Fuel Radionuclide Inventories at 1,000 

Years'^-''' 



Radionuclide'*^' 



Half-Life 
de'"" (yr) 


Curies per 

1,000 mthm'''' 


1,000 Year 

Activity 

(% of Total) 


Cumulative 

Activity 

(%) 


432 


894,500 


51.33 


51.33 


7,380 


(e) 

31,080 


1.78'^' 


53.11 


6,569 


476,900 


27.37 


80.48 


24,130 


304,700 


17.45 


97.96 


375,800 


1,755 


0.10 


98.07 


88 


967 


0.06 


98.12 


213,000 


13,030 


0.75 


98.87 


80,000 


5,150 


0.295 




100 


354 


0.020 




1,530,000 


1,933 


0.111 




15 


1,836 


0.105 




20,300 


1,240 


0.071 




5,730 


1,372 


0.079'" 




244,500 


1,984 


0.114 




2,470,000,000 


317 


0.016 




23,400,000 


271 


0.018 




2,140,000 


1,000 


0.057 




100,000 


722 


0.044 




65,000 


• 405 


0.023 




3,000,000 


345 


0.020 




90 


163 


0.009 




6,500,000 


112 


0.006 




15,700,000 


• 32 


0.0018 





241 



Am 



243 



Am 



240 



Pu 



239 



Pu 



242 



Pu 



238 



Pu 



99 



Tc 



59 



Ni 



63 



Ni 

93 

Zr 

94 , 

Nb 



234 



238 



236 



237 



126 



Np 



Sn 



79 



Se 



135 



Cs 



151 



Sm 



107 



Pd 



129 



(a) 
(b) 

(c) 

(d) 
(•> 

(f) 



Reprint of Table 2.5 of Reference 8. 

Based on ORIGEN-2 data in ORNL/TM-7431 for 33,000 MWd/MTM burnup 
PWR spent fuel, actinides plus fission products plus activation 
products. 

Radionuclides with 1,000-yr activity less than ^^^I or half-life 
less than 1 year were omitted. 

MTHM 



metric tons of heavy metals 



Includes activity of '^''^Np daughter product. 

Carb9n-14 activity may vary considerably, depending on as- 
fabricated nitrogen impurities. 



5 

Under oxidizing conditions, technetium is expected to 
exist in the (+7) valence state as the pertechnetate anion, 
which is fairly stable and highly soluble. ^'^° The 
pertechnetate anion will be reduced to lower valance-state 
compounds (TcOj, TC3O4, Tc(0H)2, Tc(OH), Tc, etc.) in a 
reducing environment. Technetium is less soluble and may be 
more strongly adsorbed on minerals in these reduced 
states. 2' "•^'''^^ Thus, if the waste environment can be made 
into a reducing environment, technetium migration away from 
the potential repository may be minimized by precipitating 
it out of solution or adsorbing it onto minerals as a direct 
result of its reduced state. Therefore, the intended end 
result of this investigation is to provide information that 
could be used to make the waste package and/or near-field 
environment reducing in case of water infiltration. This 
would reduce the rate of ''tc migration away from the 
potential repository. 

I used two approaches to conduct this investigation: a 
computational-analysis approach and an experimental 
approach. The information gained by these experiments could 
be used to tailor the waste package and/or near-field 
environment to minimize technetium migration, and to assess 
the results of the computer simulations. To my knowledge, 
no prior experimental work under the scenario laid out in 
this report has been done to assess the codes. 



6 
1.2 Computational-Analysis Approach 

The computational-analysis approach was based on 
information obtained in preliminary research. ^^'^-^ Methods 
were investigated to change the waste environment from 
oxidizing to reducing. After conducting literature 
searches, applying technical theory, setting up chemical- 
equilibrium equations, constructing porbaix diagrams, and 
interpreting those diagrams, it was determined that adding 
reducing agents to the waste environment may precipitate 
technetium from solution. ^^'^^'^^ 

The EQ3/6 software package (herein called the EQ3/6 
codes or simply EQ3/6) was used for computer 
modeling. ^^'^^'^^'^^'^' These codes were chosen because 
they satisfied seven criteria discussed in Section 2.2. One 
of those criteria is the code's development is supported by 
the Yucca Mountain Site Characterization Project.^^ 

The EQ3/6 software package has many data files, codes, 
and a data file preprocessor; however, two main codes 
characterize the geochemical behavior of aqueous systems. 
The first is EQ3NR in which speciation-solubility 
calculations are performed to determine the chemical and 
thermodynamic state of a solution. ^'^•^' The other, EQ6, is a 
reaction path code "... which models water/rock interaction 
or fluid mixing. "2' Essentially, EQ6 indicates the product 
minerals formed, differences in solution environment 
variables, concentrations of reactants in solution, gases 



7 
produced, and so on, as it incrementally adds the reactant 
to the system. 

The preliminary computational analysis was conducted 
while pursuing I was my master's degree. ^^ These 
simulations used the EQ3/6 codes to model the scenario of 
groundwater infiltration into a breached waste cask, 
followed by "Tc leaching into an environment containing a 
reducing agent. This was simulated by entering into the EQ3 
input the concentrations of the elements/compounds present 
in the groundwater and the concentration of ''Tc. The mass 
of the reducing agent investigated was entered into the EQ6 
input. The codes then calculated the effects of the 
reducing agent (metallic copper, iron, tin, manganese, and 
vanadium, and the compound stannous chloride) on 
(1) solution Eh and pH, (2) solid technetium compounds 
precipitated from the solution, and (3) the concentrations 
of technetium in J-13 well water. Well J-13 is a high- 
flowrate well near the potential repository site at Yucca 
Mountain.''^ The water composition, shown in Table 2, is 
very similar to that of the vadose and groundwater at Yucca 
Mountain and is designated a reference groundwater site.''^ 

Computer modeling indicated that the soluble TCO4" 
would be reduced to much-less-soluble compounds such as 
TC3O4, Tc(0H)2, TcOH, or Tc metal. Thus, the radioactive ''Tc 
would be replaced by innocuous ions such as Fe""* and Fe^* in 
the groundwater. The computer modeling did not indicate 



Table 2 J-13 Well Water Composition''^ 



Species 


Concentration 

(mg/L) 


HCO3" 


125.3 


Si02(aq) 


57.9 


Na* 


43.9 


SO4" 


18.7 


Ca^^ 


12.5 


NO3" 


9.6 


CI" 


6.9 


K* 


5.11 


F' 


2.2 


Mg** 


1.92 


Li" 


0.042 


Sr"" 


0.035 


Al""* 


0.012 


Fe*" 


0.006 1 


pH = 7.6 



that technetium would precipitate as TCO2. This is 
addressed in Section 4.3.1. 

Should technetium enter into solution, both the 
theoretical calculations (for constructing the Pourbaix 
diagrams, see page 6) and computer calculations indicated 
that the potential for minimizing technetium migration from 
a geological repository is promising. However, the computer 
models are limited by current knowledge limitations of 



9 

kinetic and chemical thermodynamic information. The 
possibility also exists that important low-solubility 
technetium phases were not included in the computer 
database. These limitations made experimental confirmation 
of these results necessary. 

An experimental investigation was conducted to 
determine whether placing reducing agents in contact with 
technetium-spiked solutions would reduce the oxidation state 
of technetium, precipitate it out of solution, and therefore 
decrease its concentration in the solution. Measuring 
technetium concentrations as low as indicated by the codes 
was impossible. Technetium-95m was used instead of "Tc 
because it was the most sensitive radionuclide for measuring 
technetium concentrations for the given experimental 
constraints (i.e., up to 2 months run time, budget 
constraints) . The experimental results were given as below 
detectable limits (BDL) whenever technetium concentrations 
approached those indicated by the codes. The experiments 
distinguished between a reducing agent that was able to 
reduce the technetium concentrations BDL versus those that 
had little or no lasting effect on technetium 
concentrations. This was possible even though the 
technetium concentration could not be measured as low as 
calculated by the codes. This approach for measuring the 
relative experimental effectiveness of different reducing 
agents and of evaluating differences from calculated 



10 

technetium concentrations was determined adequate by the 
Ph.D. committee and myself. 

1.3 Experimental Approach 

This research was designed to develop, plan, and 
perform experiments to evaluate the use of reducing agents 
to impede the rate of technetium migration. This 
information could then be used to tailor the spent-fuel 
waste packages and/or near-field environment to mitigate 
technetium migration should a waste cask breach and 
groundwater contact occur. In a geologic repository, this 
may be accomplished by adding a reducing agent to the 
engineered barrier system. The computer calculations 
indicated that the reducing agents (entered as the reactants 
in the simulations) would cause technetium to precipitate in 
a J-13 solution environment. This led to an associated set 
of objectives of comparing the experimental and 
calculational results to assess the results of the EQ3/6 
codes. 

The experiments consisted of reacting solutions with 
known technetium concentrations with different reducing 
agents, such as iron powder. After the technetium-spiked 
solutions were allowed to react with the reducing agent, 
effluent samples were collected and analyzed to determine 
the extent to which the reactant reduced the technetium 
concentration. Some of the experiments used flow-through 



11 

systems that pumped the solution through reaction columns 
containing the reducing agent. Other experiments allowed 
the solution to remain in semi-static contact with the 
reducing agent. The details of the experiments are 
presented in Section 3.4. Experimental results and 
conclusions are discussed in Section 4. 

1.4 Laboratory Closure 

Experimental work in the 325 Building at Pacific 
Northwest National Laboratory (PNNL) could not be conducted 
from April 1994 through February 1995 because laboratory 
activities were shut down for upgrading various safety- 
related systems. Afterwards, a phased-in restart of the 
individual laboratories took place. All those who had 
laboratories in the 325 Building were involved in a massive 
laboratory clean-up, taking numerous training classes, 
writing Technical Work Documents, and preparing Start-Work 
Packages. During this time, work pursuant to my Ph.D. 
degree was also undertaken. Several samples from one of the 
experiments were analyzed for iron concentration, a Fortran 
program was written to manipulate the data output from the 
EQ3/6 codes, and numerous EQ3/6 simulations were conducted. 
Pertinent EQ3/5 simulations are discussed in Section 2.3. 



12 
1.5 Development of This Dissertation 

Section 2 discusses both analysis and results of the 
computer simulations. Section 3 describes the preparation 
for the experiments and the actual experiments. Section 4 
discusses the results and individual conclusions for each 
reducing agent. The overall summary and conclusions are 
presented in Section 5, and Section 6 lists recommendations. 



.': i-r 









CHAPTER 2 
COMPUTATION ANALYSIS AND RESULTS 

2.1 Modeling the System 

The theoretical basis for choosing a code or set of 
codes to model the system was based on the equations that 
define what is chemically and thermodynamically occurring in 
the system.^ The system to be modeled was to use reactants 
to reduce the solution environment and to cause the 
technetium in solution to precipitate. 

For the solution to become reducing, the reactant must 
be oxidized by losing electrons. For example, oxidation of 
iron(O) to iron(II) results in the loss of two electrons 

Fe° ^ Fe** + 2e" (1) 

This expression represents only half of the reaction 
involved. There must be a corresponding reduction in which 
a reactant gains electrons and is transformed to a lower 
oxidation state. In the current case, it is anticipated 
that Tc(VII) will be reduced, for example to Tc(II) 

,. . Tc*'' + 5e" «* rc*2 (2) 



^The theory behind the equations in this section can be 
found in References 17-19, or any good chemical 
thermodynamics book. 



13 



14 

These types of reactions are called electrochemical (or 
redox) reactions because electrons are being transferred. 
Of course, many other electrochemical reactions can occur 
because of all the chemical species involved in the J-13 
solution. 

Chemical ions can undergo other types of chemical 
reactions without undergoing oxidation or reduction. For 
example iron(III) combines with water to form iron(III) 
hydroxide and, on a much longer time scale, will then form 
hematite according to the reactions 

Fe"* + 3F2O '^ Fe(OH)^ + 3H^ . (3) 

2Fe(OH)^^ Fep^ + ^Hp (4) 

However, the interaction of iron and water is not limited to 
hematite; other chemical reactions can form different 
products, such as magnetite, Fe304. In other words, each 
chemical species can be oxidized, reduced, and/or combined 
with other chemical species to form many different 
compounds. These compounds can be either aqueous, gaseous, 
or solids. 

Therefore, to model the system that uses reactants to 
reduce the solution environment and cause the technetium in 
solution to precipitate, numerous equations must be solved. 
The equations that were required for modeling this system 
will be discussed. These equations are required for each 



15 

chemical species, ion pair, and ion complex possible for the 
system. 

The Nernst Equation, see Equation (5) , was used to 
carry out the calculations for electrochemical reactions, 
such as those in Equation (1) 

E,= E^ - ^-^^y^ log K (5) 

where R is the universal gas constant, T is the temperature 
in Kelvin, n is the number of electrons being transferred, F 
is Faraday's constant, K is the equilibrium constant, and 
Ec° is the standard potential for the reaction being 
considered. At equilibrium, E^ is equal to 0, and the Nernst 
equation becomes 

0^ 2.303 i?r 

nF 

The standard potential, E^° , is obtained by manipulating the 
following standard-state Gibbs free energy change for a 
reaction, AGp" 



^Gr = -2.303 RT In K = -nFEr (7) 



to obtain 







-nf 
AGr° is determined by 



E° = ^ (8) 



. 16 

^ Gf = y ^ Gf (products) ~ ^ Gf (reactants) \^' 

where AGf° is the Gibbs free energy of formation. The AGf° 
values can be found in chemical thermodynamic tables.^'' 

Equation (7) provides a basis for choosing potential 
reactants for reducing TCO4' . The standard potential for a 
given reaction can be determined by considering the 
appropriate half-cell potentials. For example, the 
half-cell reduction potentials for Fe** and TCO4" are as 
follows: 

Fe^* + 2e" ^Fe E° = -0.44 V (10) 

TcOi, + 4if* + 3e-'-Tc02 +2H2O E° = +0.70 V (H) 

The standard potential for the reduction of Tc04' with iron 
metal is given by 

2rc04" + 3Fe + 8H' ^ 3Fe** + 2TCO2 + ^E^O (12) 

E° = 0.44 V + 0.70 V = 1.14 V 
Because E° is a positive value for this reaction, the free 
energy of formation, as given by Equation (7) , is negative. 
Thus, the reduction of TcO^" with Fe metal is 
thermodynamically favorable. Similar reactions can be 
derived for copper, tin, and stannous chloride, which showed 
favorable results for reducing TcO^" . 

The equilibrium constant in Equation (5) is defined as 
where a is the thermodynamic activity. As will be discussed 
momentarily, the activity is a parameter that also must be 



17 



p. _ L ^product 1 J (-"product 2 J MT^ 

•^reaction " -f- f- rr T •'' ^^^1 

L'^reactant 1 J L'^'reactant 2 J 



determined. 

There are additional equations for chemical reactions. 
For each ion in solution, there is an associated mass 
balance equation that calculates the total contribution for 
that ion regardless of the chemical species. For the case 
of Fe(II), the mass-balance equation in terms of molalities 
would be as follows: 

%.Fe" = ^Fe- ^ ™Fe(0H)2(aq)-^ ™Fe(0H)3- ^ ^FeC03(aq)^ ' ' ' ^^^) 

The concentrations of ion pairs and complexes are 
governed by thermodynamic equilibrium. These are 
represented by mass-action equations for the dissociation of 
the ion pair/complex. For example, the dissociation of iron 
carbonate is 

FeCO:^^^^^ ^ Fe- + COi^ (15) 

and the mass-action equation for this dissociation is the 
equation for the equilibrium constant: 



_ [ape..] [aco^3 

^FeCOscaqf rz ^ (16) 

L^FeC03(aq)J 



The thermodynamic activity, a, for each species must be 
calculated. The thermodynamic activity of species i, a,-, is 



18 

related to the molal concentration, iiii , and the activity 
coefficient, Aj, by the equation 

aj = ro,- A.,- (17) 

To solve Equation (17) , it is necessary to calculate the 
activity coefficient, which is determined by the Extended 
Debye-Hiickel equation: 

log k, = -^>,io (^i)' V^ (18) 
1 + \[l 

where A^^jo is the Debye-Hiickel parameter in terms of the 

base ten logarithm, I is the ionic strength, and z,- is the 

electrical charge for species i. The ionic strength is 

defined as 



^ = ^ 5^ ^iZi^ (19) 



2 , 



where the summation is over all solute species. 

To accurately model the thermodynamic state of an 
aqueous solution, all the above equations must be solved for 
all the species involved. The concentrations of species 
initially used to make the solution are known; however, the 
concentrations of all the other possible chemical species 
must be determined. This poses a dilemma. To calculate the 
concentrations of these other species, the mass-action 
equations must be evaluated. This requires the activity 
coefficients of all the species. To compute the activity 



. 19 

coefficients of all the species, the concentration of all 
species must be known, which was the original goal of the 
calculation. This dilemma is often solved by making an 
assumption for the activity coefficients and then 
calculating the others. To accurately solve the problem, 
iterative technigues are used. In one such technique, the 
activity coefficients are estimated, and the mass action 
equations are solved to obtain the concentrations of all 
species. The activity coefficients are then calculated 
using the new concentrations. This type of iterative 
process is continued until the calculated activity 
coefficients converge. 

• - Put the above dilemma aside for a moment and assume 
that all the solution variables are calculated. The 
interactions that can occur between the solution and the 
added reactant also must be determined. As the reactant is 
added to the solution, the number of moles of solution 
changes, the solution composition changes, and the solution 
may become saturated with new minerals. 

Adding the reactant also moves the system from 
equilibrium because the reactant dissolves into the system. 
The system is now in a state of disequilibrium and will be 
thermodynamically driven toward a state of equilibrium. If 
the aqueous solution is undersaturated with respect to a 
mineral, the thermodynamic driving force is for the 
dissolution of the compound (s) containing that mineral. The 



20 

system is driven until no more reactant is left, or the 
solution is saturated. 

To determine the state of the system, all the preceding 
equations must be evaluated in steps for each new species 
formed. For each species formed, new equations also must be 
added. If the species is a pure phase (for example, 
technetium metal) , then only one mass-action expression is 
added for the corresponding solubility equilibrium, and one 
unknown is added (the number of moles of technetium) . If 
the compound is composed of more than one endmember, then a 
new equation is added for each endmember, and the number of 
unknowns is increased by one for each endmember (number of 
moles of each endmember). 

Once again, to accurately model the thermodynamic state 
of the entire system, all the above equations must be solved 
for all the species involved. Although all the above 
equations are algebraic, and again are solved iteratively by 
similar methods, the solution is complex. Numerical methods 
using computers are required for solving the thermodynamic 
state of the system. 

2.2 The EQ3/6 Computer Codes 

Using computers for the geochemical modeling of an 
aqueous system is the only feasible means of solving all the 
equations mentioned in Section 2.1 and other related 
equations. There is not a globally adequate set of 



21 

geochemical modeling codes currently available because none 
provide for all the physical and chemical phenomena that can 
occur. Also, not all the thermodynamic data are available 
for all the species. 

The EQ3/6 software package was chosen to conduct the 
computational-analysis for this research. ^^'^^'^' The EQ3/6 
codes were chosen for modeling the system because (1) they 
could perform speciation-solubility calculations of the 
aqueous system, (2) they could perform reaction-path 
calculations of the system, (3) the supporting databases are 
extensive, (4) the speciation modeling is not hard-coded,'' 
(5) they include balance equations for computing changes in 
the amount of solvent water, (6) they are capable of 
handling some of the experimental systems used, and 
(7) their development has been supported by the Yucca 
Mountain Site Characterization Project. ^'^'^^'^' 

The EQ3/6 Software Package consists of many parts. The 
EQ3NR and EQ6 codes characterize the geochemical behavior of 
aqueous systems. There is a data file preprocessor (EQPT) 
and a software library (EQLIB) . There are also five 
external supporting thermodynamic data files. 

EQ3NR is a speciation-solubility code that is useful 
for "analyzing groundwater chemistry data, calculating 



b 



Computer codes that have data entered directly into the 
program are called hard coded. The EQ3/6 codes, however, 
have subroutines that look to an external file for the data. 



22 

solubility limits, determining whether certain reactions are 
in states of partial equilibrium or disequilibrium."^' In 
other words, it calculates the chemical and thermodynamic 
state of a solution. ^^'^' It is required to initialize the 
EQ6 calculations. - • . - 

EQ6 is the code that determines what occurs to the 
system as the reactant is incrementally added. ^'' This is 
, called the reaction path of the system. The reaction path 
calculations for this code include those for gross 
composition changes that affect the equilibrium of the 
system. They also include elements of simple 
disequilibrium, and they can be defined by rate laws for the 
addition of reactants and/or the formation of product 
phases. Appendix A discusses the description of the codes 
in greater detail. 

2.3 EQ3/6 Simulations 

Numerous EQ3/6 simulations were conducted for all the 
reducing agents. Some of these simulations were conducted 
during the preliminary computational-analysis (see Section 
1.2). However, the majority of the simulations have been 
conducted during and after interning at PNNL. A few of the 
significant simulations are presented here to help 
understand the results when the experiments are compared 
with the computer simulations later. 



23 
2.3.1 Preliminary EQ3/6 Simulations 

Figures 1 through 5 show the solution environment and 
the numerous technetium products that are indicated to 
precipitate for the preliminary simulations using copper and 
stannous chloride reactants, each in J-13 well water with 
technetium metal. Technetium metal was used in these 
simulations because the metal will exist at the proposed 
repository. Numerous technetium compounds were indicated to 
precipitate due to the very low Eh values obtained as the 
technetium was oxidized from the (0) valance state to the 
(VII) valance state. 

In the simulations that represent the experiments, 
TCO4" was entered as part of the solution rather than Tc 
metal entered as a reactant because pertechnetate solutions 
were used in the experiments. Not all the technetium 
compounds were indicated to precipitate in these latter 
simulations because the metal reactants used could only be 
oxidized to the (III) valance state and therefore, the Eh 
values were much higher than the previous simulations. The 
simulation ended once the system was saturated with the 
reactant and before all the technetium compounds 
precipitated."^ 



When a system is saturated with a reactant, it is in 
thermodynamic equilibrium, and any additional reactant added 
will have no effect on the equilibrium of the system. 



24 
Figure 1 Curve 1 shows the change in pH (left axis of 
ordinates) versus the calculated reaction progress, zi, for 
the simulation of technetium and copper metal in a solution 
of J-13. Figure 1 Curve 2 shows the Eh (right axis of 
ordinates) versus zi for the same simulation. EQ3/6 
represents reaction progress by zi, where zi represents the 
integral number of moles of initial reactant that has 
dissolved. Zi is not necessarily a linear function of time. 
In these EQ3/6 simulations, the reactant used for the 
reaction progress was technetium metal. Zi does not 
represent solution concentrations of technetium since some 
dissolved technetium was indicated to precipitate as TC3O4, 
Tc(0H)2, Tc(OH) and Tc, as will be shown in Figure 2. 

Figure 2 Curves 1, 2, 3, and 4 show the number of moles 
of TC3O4, Tc(0H)2, TcOH, and technetium metal that are 
indicated to precipitate from solution as the reaction 
progresses, respectively. In these simulations, once a 
compound was indicated to precipitate, the compound did not 
dissolve and re-enter into solution, nor did it convert to 
another compound. Figure 3 is a combination of Figure 1 and 
Figure 2 . The technetium compounds that were indicated to 
precipitate from solution did so as the solution environment 
became reducing, i.e., as the solution redox potential. Eh, 
decreased (Figure 3) . 

Figure 4 and Figure 5 show similar plots for the 
stannous chloride, technetium metal, and J-13 well water 



-i*s V •»v* 



'%,^t,^'\. 



K -,; 



If- 



. V %. 



25 



Eh(V) 



O 05 



CO 
O 



d 



d 



d 



d 



CO 

d 



d d 



o 
d 



T- CM 

d d 



CO 

d 



d 



IT) 

d 



d 




^ 



" " ' " ' ^' ' i , , . . I . ■! , ■ ■ ; — i — ; ■ ■ ■ ■ ; — ; — ; — ; — I ^ 

CDLO^COC^^OCnoOr^CDlDTj-mtM^O^ 



d 



00 

d 



O) o 
d •^ 



■ LU 



W 

<? §> 

c 
o 

o 



UJ 



£ 

3 
M 
CO 
0) 

E 



w 
N 

N 

S2 

> 

(U 

E 
c 
o 

> 

c: 
UJ 

c 
o 

••— ' 

o 



CO 

ra 

4^ 
<o 

E 
E 

c 
o 

c 

o 

w 

_a> 

o 

E 



03 
> 





E 

3 



0) 






> 

c 
HI 


c 

0) 

w 






c 


n 


o 


(U 






3 


w 


o 


■^ 


CD 


u 




c 


c 


on 


o 


(n 


% 


(A 


-1 




E 


o 






w 


Q. 


8 


o 


Q. 


o 


O 


m 


() 


Q) 






CD 


(D 


CO 


n 


O 




UJ 


o 



Hd 



3 



26 






H O 






O 5 

•^ o 



M- 



«MMM^B III ij 



O 
■f 
UJ 



M 
M 

o> 
p 



c 
o 

■£3 
a 

a. 



*■ ^"" ^ """ ^ """ *■ " ■■■ ^ "" i ^ "ii ^ ^"•' t. L L i- : — 



Ui 



LUliJliJLlJLjJuJuJ|jj'7'7';-Tp»^ 



0> 



■§ 




Q) 




■** 




«•• 




O 




c 




o 
































n 




a> 






ir\ 


TO 

-J 




o 




TO 


Q 


O 


Q. 


TI 


C 

o 


C 


ri 


3 


m 


o 


11 


Q. 


i 


F 


0) 


o 


£ 


O 




E 

3 


o 


^_f 


T 


(U 


Crt 


c 


01 


sz 


(1) 


o 

0) 


E 


1- 


(U 


■o 


£ 


(i> 








TO 


■~* 






Q. 


N 


O 




£ 


N 


Q. 


O) 


(- 


c 


o 


CO 




TO 


TO 


<11 


3 


() 


fc 


c 


to 


f 






0) 


^ 


n 




Q. 

o 


T3 


O 


3 


<o 


o 




u. 


CO 

O 


E 



pnpojd Luniiaugoai pajBijdpajd jo sa|0|/\| 



LU o 



CM 



Moles of Precipitated Technetium Product 



27 



+ 

lU 



9 

Ul 



III 



<9 

Ul 



Ul 



<M 
lU 




(f) 

o 



c 
o 

•s 

(0 

q: 



(a) 43 puB Hd 



n 

c 
a 

I 
o. 

UJ 
(0 

•o 
c 

3 
& 

E 
o 
O 

E 

3 



»- to 

O O 



El 

CO (/) 

c c 

F O 

o E 
.i: O 
> O 

!| 

O (U 

IE 

o o 

II 

^ o 

11 

o 5J 
HI o 



CO 

£ 

3 



28 



Eh(V) 



oo 
d 



o 



d 



d 



o 

d 



CM 

d 



o 



m--^-^-^-^ 



I 

I 

it 
ii 

il 

ii 

•r 
il 
il 
il 

il 

il 

■1 
i I 

i I 

i I 

i i . 



°- i 
o 




p 



o 

TO 
0) 



Ill 

o 



I 

Ul 

o 



CO 



N 


E 






3 










N 


fl) 




(A 


«- 




3 


(- 




!" 


o 




0} 

> 


CD 

■4-' 




*-• 


CD 




c 






<u 

F 


c 




o 


o 






(/) 




> 


(U 




c 


o 




m 


F 




c 






u 


o 




3 


CD 




o 


n 




w 


E 




<^- 


-I 




o 


c 




V) 


^ 




o 


Si 

O) 




■4-' 

CO 


0) 






c 




eo 






> 


CD 




(D 






E 

c 


a> 




o 


r 






QJ 




> 


to 




c 


9^ 




UJ 


Q. 




o 


£ 




•*-< 


CO 




-I 








■n 




o 


r 




W 


CD 




c 


V) 




o 


CO 




03 






3 


o 




E 


Q. 




(/) 


c 




0) 


o 




■a 


V, 


•o 


I— 


in 


Q) 


o 


(D 


> 


f 




O 


O 


(- 


^ 


3 




■a 


O 


o 


CO 


{- 


2^ 


CD 


c 


x: 


CO 


3 


^- 




(/) 


CD 


(J) 


CD 


£ 


CD 


E 




CO 


CD 


o 


(U 


(U 


UJ 




E 



Hd 



& 

3 

u. 



29 




w 

O) 

o 



o 
o 

CD 



05 



T3 

C 

Q. 



Q. 

O i5 

Q. C 

■D "> 

to a) 

O 



o 



(3- 

c 

3 TO 
O <1> "O 
o >- 0) 



o 

E 

3 



o 



in 
tn 



o 



w 

CO 



o 

0) 



E £ 



■ ^2 
oj 10 c 

CO i- 



E 

3 



Q. 

c 
o 



.E o 

^ CO -^ 

TO 2 15 

3 o s 

E .= .E 

W £ >5 

0^ 5 w 

°? i 

-C 3 = 
O O >*r 

^E ° 

3 C: m 

^ c= ^ 
TO E 3 

55. 2 f 

to ^ 2 

« ^ S? 
O o 52 
HI OJ c 



S|Bjau!|/\| pnpojd LunijauLioaijo S8|0|/\| 



S 

Ur. 



30 
simulation. The reactant used for the reaction progress was 
technetium metal. Once again, zi does not represent 
solution concentrations of technetium because some 
technetium is indicated to precipitate. Figure 4 Curve 1 
shows the major difference is that the solution environment 
became acidic, not alkaline, as in the case for copper 
(Figure 1 Curve 1) . 

2.3.2 Interpreting the EQ3/6 Figures 

Figure 6 is a symbolic interpretation of the computer 
simulation results. A discussion of this figure is followed 
by a detailed explanation of a figure similar to those in 
the remainder of this dissertation. This section can then 
be referred to as the graphs become more complicated. 

Figure 6 shows the symbolic interpretation for the iron 
product minerals formed in the simulation that used the 
reducing-agent iron as a reactant with a solution containing 
aqueous technetium as pertechnetate (TcO^") . This is a 
representation of a fluid-centered flow-through open system. 
The changes that occur to the system are from the point of 
view of the packet of water as it progresses through the 
reactant. The packet of water that was being followed in 
the simulation is represented in gray. As the simulation 
began, the spiked J-13 was highly oxidizing. The iron was 
oxidized to hematite (Fe203) and nontronite-Ca 
(Cao.i65Fe2Alo.33Si3_67H20i2) , essentially iron (III) oxides. As 



/ 


ff. 




i£ 


} 



0) 





31 



© 




c 
'n 

x 
o 

t5 

CB 
Q) 

d 

C 

o 

0) 

a: 



O) 

c 
"n 

x 

o 

(0 
CO 

c 
o 



03 

c 

N 

■a 
x 

o 

< 

c 
o 

OS 

cc 



o 

O) 

c 
'c 
c 

O) <u 
0) -c: 

^:^ 

r 8 

CD ^ 

CO -^ 

8l 

11 

(D -^ 
DC CO 

>^ 

CO 

^^ 

c iS 

— CO 

ii 

x: CO 

CO -g 

CO x 

'Z O 

0) :=> 

to =, 

$ V 

H- LL 

O CD 

o o 
_ CO 



O 

c CO" 
O O 

a: c 
o 

t! 

•6)2 

O CO 
CO Q 



U- D) *-: 

III 
o k 05 

>^ CD c 



^ o 2 

O CO 



X 

o 

CO 



CD 



Z2 

E 



^ E 

>s (D 

CO N 
C XJ 
<D X 
Q. O 

O CO 

x: 'c 

2"o 



CD 

C 

o 

<D 



::r- co 
^^co 

CO 

its 

CO E 

- CD 
D) ^ 
C O 

N (D 

. X ^ 
D) O - 

- ^ ^', 
"D 0) O 

J$ CO ^ 

O — Q) 



■D 

c 
CO 



c 
o 



o 

LL 

TD 

C 
CD 
O 
I 

•o 

'l3 



a> 



:£ ^ :5 O 



CO 

<D 

E 
o 
o 

CD 
J3 



O 
CO 

CD 



CD 

SI 



O 

o 

CO 



c 
g 

E 

CO 

CD 

t vi"* [3 ;^ 

i: i_ .i= CD 



^- CO < 

c3 o Ir 

CO "D ^ 

<D -= c: 



32 

the packet moved forward through the iron reactant, the 
water became less and less oxidizing because the oxygen 
available to this packet was being used to oxidize the iron 
and precipitate the compounds. Eventually, it was not 
oxidizing enough to precipitate hematite or nontronite-Ca. 
It was able, however, to precipitate the iron(II) oxides 
pyrite (FeSj) and cronstedite-7A (Fe2Fe2Si05 (OH)^) . The codes 
then terminated the simulation because the system reached a 
point where nothing was changing because all the available 
dissolved oxygen was exhausted. 

Figure 7 shows the technetium concentration and solid 
product minerals that were calculated by the same computer 
simulation. They are plotted versus the reaction progress 
(zi) . Zi is the measure of the reaction progress and is 
represented by the integral number of moles of initial 
reactant that dissolved. In this simulation, the initial 
reactant was iron metal. 

Curve 1 (upper solid line) represents the technetium 
concentration in parts per billion (ppb) . This is the only 
set of concentrations in this figure. All the other lines 
represent the number of moles of solid products 
precipitated. 

The first solid iron products indicated to form by the 
simulation were hematite (represented by Curve 2) and 
nontronite-Ca (represented by Curve 3). Nontronite-Ca 
formed until there was a reaction progress of 3.1 x lO"^. 



33 









1 — 
1 


JV 




B*^ 


"JL 


on and 
number of 










/ . 


, - 




■ 8 


_^ 




in 
1 


1 
1 






/ ' 






d 


n of 
egra 


















•« 




o 


1 
f 






/ ' 








tati( 
e in 




I 


1 


^ 


/ 








Q-5 




O 




tt 


/ 1 








J^ >^ 




o 


\ 


1 


/ 








CLT3 




1 

LI. 


1 
1 


i 


: / 1 ! 
> 1 3 ■ 






1 


lated 
sente 




u. 


1 


a 


/ "1 






d 


3 (D 
1-- 




< 


1 


t 


1 1 _ 








■^ Q. 




"> 


jf" 


■ 1 ^ ' 








CO m 




£ 


1 


EL 


1 ^ ' 








^ 




1 


1 
I 




/ 1 

T *~ • 












1 
2 
o 


1 

1! 




1 : 








eacta 
ssan 








• S" ~ / 1 






rM 


a: £ 






3 1 

11 




Tc (ppl 
(Curve 


% 


/ 


d 

N 


r Iron 
iprog 








■ 1 


/^ 






sfo 
ctioi 






!bl 




i 


/ 


'■^ 








s ( 




T 






E 


CD CO 






e 1 




1 






O) 


I- as 
0) S 






w 


J 


■ 1 






2 


C m 






1 
1 




J 






015 

tion P 


duct Ml 

re of thi 






1 




1 ^ 















1 ' 






d CO 


P 3 






1 




■ / ° 






CD 

a: 


0- CO 






1 




, / 








1^ 






r 


CO 










.2 (o . 








/ 


1 

.- 8 

9; ; 


i">>^\j<>'_ . ■ 




■ 


8 

d 


oncentrat 
ng zi. Zi i 
dissolved 










I 

& . 






. 




urn C 
reasii 
thas 










s , 


'<■ 








■•= y CD 












in (ppb) 

Ca 

B-7A 








echne 
with in 
tant th 








1 ' 


5 : 


Tc Free Rl 

Hematite 

Nontronite- 

Cronstedtlti 

Pyrite 

Tc304 


X 




1- 


- 


d 


ulation T 
1 pounds ' 
itial read 






1 


l| 


1 ; 


1:1. 


1 


. 




E = = 
.E — 










! 
z . 


"lit? 


1 


■ 




3/6 S 
perc 
esof 






i 


'\ 








■ 




Co? 
UJ 8 E 


u 


■III 1 >ii ■nil 




^ 






ro 







8 

+ 


s ^ 


s 


? 9 S 


9 9 9°^ 


CM 




UJ 


LU LU 


111 


HI UJ LU 


UI LU UJ ull u!l 


uJ 


uJ 








(saioLu) 


S|BJ 


auji 


/\l }onpoj< 


J pUB u 


DiiBJiuaouon fadd) mniiamnpt 








3 




34 

It was at this point that the solid iron (II) compounds were 
the thermodynamically favored iron compounds to form. The 
iron (II) compounds indicated to precipitate were pyrite 
(represented by Curve 4) and cronstedite-7A (represented by 
Curve 5) . All three iron compounds, hematite, pyrite, and 
cronstedite-7A, formed until the simulation terminated. The 
first technetium compound indicated to form at a reaction 
progress of 1,3 x lo'^ was TC3O4 (represented by Curve 6). 
The technetium concentration. Curve 1, decreased as expected 
once TC3O4 (Curve 6) began to precipitate. Actually, the 
graph shows that the concentration started decreasing at the 
data point before TC3O4 because the curves are plotted as 
straight lines between calculated data points. Had a data 
point existed at a reaction progress of 0.00128, the drop in 
the technetium concentration would have started at that data 
point, not at a reaction progress of 0.001. As the reaction 
progressed further, the iron reduced the technetium to 
Tc(0H)2 (Curve 7), beginning at a reaction progress of 
1.8 X 10"^, and it continued to form until the simulation 
terminated. 

As the computer simulations are compared to the 
experiments in later sections of this report, the data will 
always be plotted versus reaction progress on the upper axis 
and time on the lower axis. it is important to understand 
that the two are not related. The reaction progress, 
although plotted on the same figures as the experiments, 



35 

could actually be substantially longer or shorter than the 
experimental time frame, i.e., months or seconds instead of 
days . . Y 



r 



2.3.3 Simulations Modeling the Experiments 

From 1994-1995, during the time of the 325 Building 
stop-work order, numerous EQ3/6 computer simulations were 
conducted in an effort to model the experiments more closely 
than the first simulations (presented in Section 2.3.1), 
which were conducted by the author in 1989. ^^ For instance, 
rather than using technetium metal as a reactant in the 
simulation, TCO4" entered into the codes at the same 
concentrations as the experiments with the same quantity of 
reactants as the experiments. Updated computer simulations 
for all the reactants have been conducted because the codes 
have also been modified^' since 1989. The solution used in 
the modeling was J-13 well water, Table 2. The composition 
of J-13 well water was taken directly from the Site 
Characterization Plan,!^ except that silicon was converted 
to silicon dioxide. Strontium, which was obtained from work 
conducted at Lawrence Livermore National Laboratory, was 
also added. ^^ 

A simulation. Free Run, used the same initial amount of 
technetium and iron as those used in one of the iron powder 
experiments. The technetium and iron concentrations versus 
reaction progress obtained from this simulation are 



36 



% 



a 



a 







9 



0) £ 



-> 9 






d 






CO 

8 

o 



15 



o-e - ♦- 



n 




■a- a 



II 



a. a. 

a. Q. 

c c 

01 a: 



E 

Q. 
Q. 

C c 

ce a: 



E E 



u 0) u a> 



r--^. 







■uu-Lj — *■ FY ^ *■■"■ ■ ■ -—•• 



CSI 

8 

d 



w 
o 



c 
o 



8 rs 



CO 



in 
d 



8 o 



CM CO ^ LO (D r^ 

p o o o o o 



LU LU UJ LU 



suoijBjjuaouoQ (Ludd) uoj| pue (qdd) Lunijeuijoai 



g 
"S 

a? 



3 
(O 
CO 
0) 

E 



in 
p 



c 
o 

■s 

CD 
O 

o 



(U li. 



v> 
N 



N 

c 

10 
CO 



■o 

> 
o 

CO 



I^ CO 

o c 



c 

(0 

c3 

CO 



CD '^ 



a> 
o 
c 
o 
o 



o 

c 

CD 



0) 

c 
o 



> 

0) 
Q. 
W 



0) CO 



CD tz 



3 
O 

CD 

O 

CO 

o 



^ CD 



TO 

3 

E 
w 

CD 
CO 

O 
LU 

c 
o 



^ a; 

(U Q. 
W TO 



s: o 

TO ™ 
to (« 

E 3 
Q.O 



00 

£ 

s 



j^-ii .^.„ 37 



indicated by Figure 8 Curves 1 and 2, respectively. The 
results obtained from the codes still indicated that iron 
should reduce the technetium concentration orders of 
magnitude BDL. 

Because the experimental technetium concentrations were 
not reduced to the values calculated by the codes (as will 
be discussed in Section 4.3.1), simulations were conducted 
having various iron and technetium compounds suppressed from 
forming to determine if there was a solubility controlling 
compound that prevented the measured technetium 
concentration from decreasing to the values indicated by the 
codes. In all of these simulations, the technetium 
concentrations were orders of magnitude below that which was 
experimentally observed. No solubility-controlling solid 
technetium compound was identified because the experimental 
and calculated results did not agree. 

Attempts were also made to model the code input file to 
be closer to that of the experiments, i.e., a solid-centered 
flow-through open system, by using back-to-back simulations 
with an altered EQ6 pickup file (which is also an EQ6 input 
file) . The pickup file was altered to include fresh Tc- 
spiked solution. By having back-to-back simulations with 
fresh Tc-spiked solutions, the new technetium solutions 
would be added to the same initial amount of iron, similar 
to the experiments. However, none of these simulations 
accurately modeled the flow-through experiments. 



38 

A limiting case simulation, Limited Run, was conducted 
in which the initial amount of iron entered was just enough 
to reduce the final technetium concentration to that 
obtained in the experiments. The technetium and iron 
concentrations versus reaction progress obtained from this 
simulation are indicated by Figure 8 Curves 3 and 4, 
respectively. According to the results of Limited Run, a 
technetium concentration of 1 x lO"'' ppb. Figure 8 Curve 3, 
corresponds to an iron concentration of 0.009 ppm. Figure 8 
Curve 4. This would be BDL for the iron analysis conducted 
in this research. However, as will be discussed in 
Section 4.3.1, the iron concentrations obtained 
experimentally were higher than calculated by the codes. 
The reason for these discrepancies are discussed in 
Section 2.4. 

Although EQ3/6 cannot accurately model a solid-centered 
flow-through open system, it can model a closed-system 
experiment where the solution remains in contact with the 
reactant. This is similar to the Sealed Vessel experiments. 
A computer simulation of a closed system was therefore 
conducted. The technetium concentrations from the EQ3/6 
simulation of the closed system are plotted versus the 
reaction progress in Figure 9. The pH and Eh from this same 
EQ3/6 simulation are plotted versus the reaction progress 
and are indicated by Figure 10 Curves 1 and 2 , respectively. 
The pH corresponds to the left axis of ordinates and the Eh 



^^■fi; 



-::<■ \ » \ ?■ 



■-■t r-'.: %■' 



39 



II 




X 

O 
o 



O 

O 



o 
o 
d 



CO 
o 
o 
d 



CN 


^~«> 


o 


N 


O 




o 


</) 




(rt 




(1) 








O) 




O 








□. 




c 




o 






IT) 


o 




ro 


X 


(U 


d 


q: 



s 

d 



J — liiiii tj " — mil I I — kuiiu-i — 1 — - > ■- ^ -"■■ 



in 

8 

o 
d 



UJlLluJlilLlJlJJUJLlJLjJuJuJuJuj 



(saioLu) spnpojd pajeiidpajd '(qdd) uo!}bj}U90uoo Lunuauiioai 



(0 

s 



o 

<D 

£ 

°l 
§1 

(U CO 
E ■= 



— CD 

■ CO 
E £ 

0) _ 
w 5 

W E 

■o i^ 
0) o 

O o 
o E 

0) S 

(D _ 

d: w 

li 

JO ^ 

E oj 

W ^ 

' (D 
to CO 

?5 £ 



UJ 

(3) 

0) 



<a 



40 



Eh(mV) 



o 


o 


o 


o 


Q 


Q 


Q 


O 




o 






^ 






o 


CO 


h- 


CD 


lO 


CO 


CM 






•a 

c 

a. 

0) 



■o 

c 
CO 

w 

£ 

o 



c 
o 

CD 

<u 



o 

(0 

CD -O 
(U 0) 

E > 

£ w 

w ^ 

"Z w 

N ™ 

(U cS 
|£ 

d 2 

•.^ CD 

3 — 

E .5 

w ;e 
?| 

<D E 

■a ^_ 
5 o 

o ._ 

§1 

i 3 

^5 

^ S? 

>s 0) 

w ^ 



Hd 



& 

3 
O) 



41 

to the right axis of ordinates. The technetium 
concentration dropped 8 orders of magnitude as the solution 
environment became reducing. As will be discussed in 
Section 4.3.2.1, these simulations were consistent with the 
experimental results when iron was the reactant. 

2.4 EQ3/6 and the Experimental Results 

Reasons for lack of agreement between the measured and 
calculated activities were sought, such as (1) the solid 
technetium compound that controls the solubility in the 
experiment is not listed in the database, (2) the solubility 
listed in the database is incorrect, (3) there are 
limitations in the code's modeling capability, or 
(4) limitations of the experimental setup prevent the system 
from responding as calculated. An example of the latter is 
slow reaction kinetics such that the calculated results, 
which are based on chemical thermodynamics, are never 
realized within the limited time frame of the experiment. 

The reason for the discrepancy between the experimental 
results and the computer calculations of the fluid-centered 
flow-through open-system simulations, such as Free Run and 
Limited Run, was that EQ3/6 cannot accurately model flow- 
through experiments with continuous fluid recharging (called 
a solid-centered flow-through open system) . As mentioned in 
Section 2.3.2, the computer simulation models fluid-centered 
flow-though open systems that differ in that the changes 



42 

that occur from a packet of the fluid's reference point are 
recorded as the packet travels through the reactant. The 
changes that occur to the fluid at a spacial reference point 
as new fluid is passed through the same reactant bed are not 
recorded. Because no other computer codes will satisfy the 
seven criteria stated in Section 2.2 and model these 
experiments, back-to-back simulations (mentioned on page 37) 
were conducted using an altered EQ6 pickup file (which is 
also an EQ6 input file) in hopes of modeling the flow- 
through experiments. The pickup file was altered to include 
fresh Tc-spiked solution. However, none of these 
simulations accurately modeled the flow-through experiments. 

Another suspected reason for the discrepancy between 
the computer modeling results and the experiments was that 
the technetium in the experiments was adhering to the 
reactant and not just being reduced by it. The EQ3/6 codes 
cannot model sorption. Experimental evidence, which 
suggests that sorption occurred, is presented in Section 
4.3.1. 

The results from many of the simulations where the 
various iron and technetium compounds were suppressed were 
presented to Dr. Walter Gray.^ A meeting was then set up 



d 



vr^n'r^ll^-^^'' r?""^^ ^l ^ ^^^^°'' Scientist in the Radiochemical 
Processing Group at PNNL and was the direct onsite advisor 
for this research. 



43 

with Pete McGrail/''° who is very knowledgeable regarding 
the EQ3/6 computer codes. McGrail also suspected that the 
technetium was adhering to the iron rather than being 
reduced by it. He mentioned that the codes do not model 
adsorption, and therefore, the experimental results will not 
agree with the results obtained by the codes. McGrail also 
mentioned that the codes cannot model a solid-centered flow- 
through open system, even though the code manuals indicate 
it may be possible. ^^ Therefore, the experimental results 
of flow-through systems may not agree with the results 
obtained by the codes. 

The simulations using copper and tin were also not in 
agreement with the experiments because copper and tin both 
appeared to develop passivation layers that prevented them 
from further reducing the technetium concentration once the 
layer developed. 

Another factor that may have caused disparity between 
the experimental and calculated results is slow reaction 
kinetics. The possibility exists that the calculated 
results were never realized in the limited time frame of the 
experiments. Evidence suggesting that this occurred in the 
experiments is presented in Section 4.3.2.2. 

Although the EQ3/6 codes cannot accurately model a 
solid-centered flow-through open system as in the FT-4C and 



^Pete McGrail is a Staff Engineer in the Applied Geology and 
Geochemistry Group at PNNL. ^ 



44 

Bench-top systems, the EQ3/6 codes are continually being 
updated, and it is planned to add this option as a regular 
feature into the codes in the future. ^'^ After adding this 
feature, the codes should be able to model the experiments. 



• ■ .{. 



■ ':y 






I f ■< ' .'-■•■' ♦ ■ ; ' «-f 

CHAPTER 3 
EXPERIMENTAL PREPARATION AND DESCRIPTIONS 

3.1 Technetivim Solutions 

According to the EQ3/6 computer calculations, the ''tc 
concentration in the solution can be decreased to values 
between 1 x lo'^ ppb and 1 x lO"''^ ppb. Technetiuin-99 is a 
beta emitter with a detection limit of approximately 
1 X 10"'' ppb for the system used at PNNL. Therefore, all 
the ''tc concentrations indicated by the codes would be BDL 
for beta detections. Using a different isotope of 
technetium (element that has a different mass number for the 
same number of protons) would be of benefit because they 
would behave similarly in chemical reactions, but their 
nuclear characteristics would be different. Technetium-95m 
was the isotope chosen because it is a gamma emitter, and it 
has a detection limit of approximately 1 x lO"' ppb for 
gamma spectroscopy. The difference in detection limits for 
these two isotopes is over 8 orders of magnitude. 
Therefore, if ''^"^c is used, some of the technetium 
concentrations indicated by the EQ3/6 results would be 
measurable in the experiments. 

Although the codes indicated that the technetium 
concentration would be BDL for ''tc, the initial experiments 



45 



''i^'^ 






46 



were conducted with "tc to determine if that was indeed the 
case. Technetiuin-99 was used first because it is less 
expensive than '^""Tc and because ''tc has a long half-life 
(Table 3) . For those cases where lower detection limits 
were required, as determined by the "tc experiments, 
similar '^"'Tc experiments were conducted. 



Table 3 Properties of Technetium Isotopes of Interest 



Radiation 



Half-Life 



Specific Activity- 



Detection limit 



"Tc 



60. 5 days 



8.48 X 10^" Bq/g 



1 X 10-3 ppb 



^c 



13 



213,000 years 



6.03 X 10^ Bq/g 



- 1 X 10"'' ppb - 4 X 10"^ ppb 



3.2 Technetium-Detection System 

Because the two technetium isotopes emit different 
types of radiation (see Table 3), different detection 
systems were used to analyze the samples collected from the 
experiments. Samples containing ''tc, a beta emitter, were 
analyzed using a model 2260 XL Packard Instrument Company 
Tri-Carb Liquid Scintillation Analyzer (LSA) . This system 
was chosen because the efficiency for counting beta 
particles is very high, up to 100% in ideal situations. The 
theory of how the LSA functions and sample spectra are 
described in Appendix B. 



47 
Solution samples that contained '^"'Tc, primarily a gamma 
emitter, were analyzed using both high-purity germanium 
(HPGE) and sodium iodide (NaI[Tl]) well detectors. The 
Nal(Tl) detector was used almost exclusively. It is highly 
efficient, it does not need to be cooled with liquid 
nitrogen, and it was accessible for very long count times. 
Long counting times were essential because the technetium 
concentrations could decrease BDL. The reaction column was 
analyzed using an HPGE/translational column assembly. The 
theory of how the Nal(Tl) and HPGE function is described in 
Appendix C. 

3.3 Material Acquisition 

Four reducing agents that were investigated using the 
EQ3/6 codes were purchased to experimentally determine a 
method to reduce the rate of technetium migration at the 
potential repository. They were also used to compare the 
experimental data with the results indicated by the EQ3/6 
calculations. The reactants obtained were copper shot (1 to 
4 mm diameter), copper powder (-50 mesh, i.e., < 0.297 mm), 
iron powder (-48 mesh, i.e. <0.303 mm), stannous chloride, 
and tin shot (1 to 4 mm diameter) with purity levels of 
99.0%, 99.5%, 99%, 98%, 99.8%, respectively. These 
impurities are negligible. The only impurity of concern 
would be iron because, as the experiments in this research 
show, iron was able to reduce the technetium concentration 



48 

to BDL. However the iron impurity was only 0.005%, which 
was not enough to cause the technetium concentration to drop 
BDL in the experiments (Section 4) . Also, as will be 
discussed in Section 4.3.4, the reaction kinetics of iron 
was slow compared to the time-frame of the experiments and 
the iron impurity did not affect the experiments. The 
reactant assays are given in Appendix D. 

Approximately 1 mCi (3.7 x lo'' Bq) of ''Tc was purchased 
as NaTc04 in 5.0750 mL of solution (water) from Isotope 
Productions Laboratory. The data assay from the vender, 
shown in Appendix D, showed that no radioactive impurities 
were detected in the "Tc. Approximately 1 mCi (3.7 x lo'' 
Bq) of '^'"Tc was purchased as NaTcO^ in 5-mL of solution 
(water) from Dupont Pharmaceutical. The data assay from the 
vender, shown in Appendix D, also showed that no radioactive 
impurities were detected in the '^""Tc. 

3.4 Systems Used and Experiment Descriptions 

Four experimental systems were designed, developed, and 
used to conduct the experiments to determine a method to 
reduce the rate of technetium migration in a groundwater 
environment. These four experimental systems were the FT-4C 
Flow- through System, the Bench-top Flow-through System, a 
No-flow Beaker System, and a No-flow Sealed Vessel System. 
These systems fall under two main categories: flow-through 
systems and no-flow systems. Schematic diagrams of these 






6 

CO 
tZl 



sj 3 



..-5:11 








49 



a 

u 

CO 



I 



I 

ft 

t 

o 



E 

t 

I 

I 



6 
3 







1 







50 

systems are pictured in Figure 11. The flow-through systems 
and experiments conducted with these systems are described 
in Sections 3.4.1 and 3.4.2. The no-flow systems and 
experiments conducted with these systems are described in 
Sections 3.4.3 and 3.4.4. 

The Translational Column Assembly (TCA) was also 
designed. A schematic diagram of the TCA is shown in 
Figure 12. The TCA held the reaction column 
(6-mm ID X 3 0-cm long) from the flow-through experiments in 
a specially designed leak-proof teflon cylinder. This 
cylinder was placed in a lead cave with a window. An HPGE 
detector was located outside the lead cave, in front of the 
window. The teflon cylinder was passed by the window at 
0.25-in. intervals, and the activity along the column was 
measured. The purpose of designing this TCA was to see 
where the technetium was depositing along the length of the 
reaction column. This would enable further understanding of 
what mechanism was causing the decrease in technetium 
concentration. 

The experiments shown in Table 4 were conducted to 
determine if copper, iron, stannous chloride, or tin would 
reduce the rate of technetium migration in a groundwater 
environment. The scenario chosen to be modeled by these 
experiments was waste cask breach with groundwater 
infiltration. This was simulated by allowing technetium- 
spiked solutions to react with each of the reactants under 



r--'i ■■■■ 



51 



■a 



(0 
0) 

o 

CO 



c 
E 

o 

O 

c 
o 




o 

T3 

(D 

+-» 

L_ 

en 

c 

ig 

(D 



O 



5 

o 

T3 



5 

c 
E 

O 

o 

c 
o 

rs 

CO 
0) 



CM 



•p 
u 

3 
•O 

c 
o 
o 

(0 

•p 

c 

0) 

6 
•H 
Vh 
0) 

04 



H 



System 
Used ' 




































CO 


(D 


CC 


CD 


CD 


JD 


^ 


■o 


■D 


CO 


CD 


CD 


J3 


CO 


CD 


O 


J3 


c 




CSJ 


ro 


m 


n 


^ 


-^ 


in 


in 


tn 


f^ 


n 


^ 


m 


fO 




■^ 


o 




o 


o 


o 


o 


O 


o 


o 


O 


o 


O 


o 


o 


o 


O 


o 


o 


- « _ 














' 


' 


' 


^~ 


^~ 


' 


V- 


' 


■^ 


' 


'" 


.2 -fc -Q 
*' c a 


■ 


X 


X 


X 


X 


X 


X 


X 


X 


X 


X 


X 


X 


X 


X 


X 


X 


"E ® a 




































c 
o 




OvJ 


in 


in 


in 


CD 


t 


CO 


CO 


CM 


CO 


in 


•^ 


o 


o 


CD 


Tt 




»- 


■^ 


1- 


'- 


CO 


G) 


CM 


og 


1- 


y- 


T- 


CD 


1- 


<>- 


CM 


<J> 


u 




































a> 




































■D 








































































U 

3 
C 
O 


■ 


ID 

m 


^ 

S 


o 

1- 

cn 


o 


o 

(- 

CD 


O 

1- 

S 


o 

y- 

E 
in 

CD 


o 

1- 

£ 
in 

CD 


O 

H 


o 

H 

CD 
CD 


o 

H 

CD 
CD 


in 


o 

S5 


o 

1- 

CD 
CD 


O 

1- 

CD 

ay 


tn 


TJ 




































(B 




































cc 




































High 

jreSet 

si) 






































C\J 
CSJ 


CO 

o 


CO 

o 


CvJ 

o 










CD 
CM 


CM 


(J) 

eg 




in 


CO 

o 






Low/ 

ressi 

(P 




g 


in 


? 












§ 


8 

CvJ 


(35 




CO 

o 
in 








Q. 




































flj ^ 






o 






























owrat 
L/day 


o 


CO 

00 


o 

6 
in 


o 
in 

6 


o 
in 


o 
o 


o 
o 


■ 


• 


o 
in 

in 

CM 


9 
o 
in 


o 

ci 
in 


O 
CD 

r— 


o 
in 


g 


' 


eg 
in 


C-S 






o 














o 
















a> 




































k. 




































3 








































































(B 


















O 




















in 

CM 


C\J 


CNJ 


CM 


in 
CO 


■<3- 
CM 


CM 


O 


1^ 
O 


CM 


CM 


s 


CSJ 


CM 


s 


eg 


CM 


E 


















CM 


















0) 




































1- 




































^rf 




































ctan 

ass 

ams) 








































h- 


CO 


CM 


00 


CO 


co 


CO 


in 


CO 


en 


CSJ 


in 


r^ 


o 


1^ 










o 


o 


CM 


CM 


CvJ 


O) 




00 




CD 


C35 


o 


.^ 


"si- 


' 


o 


F] 


<J> 


o 


CD 


eg 


CM 


CM 


o 


,_ 


.,_ 


,,_ 


in 


CD 




T— 




CO 


CM 


CM 


C'J 




CO 


CO 


CO 


■^ 


'J- 


'3- 


CO 


CO 


CO 




h- 






k_ 


k- 


k_ 


^ 


^ 














^ 








3= 


C 




CD 


0) 


CD 


CD 


CD 


a) 


(1) 


0) 








CD 








3 


n 




u 


■D 


U 


T3 


■n 


T) 


T) 


■n 








■n 








+-< 


u 

n 


, 


5 


s 


5 


5 


5 


% 


% 


% 


o 


o 


b 


5 


1=; 


o 




^ 




u 


u 


o 


o 


o 


n 


O 


o 


^ 


Ji. 


JZ 


o 


^ 


x: 






u. 


Q. 


CL 


a. 


Q. 


a. 


a. 


Q. 


CO 


O) 


CO 


Q- 


CO 


CO 


O 




CC 




a> 


Q) 


<D 


OJ 


CD 


CD 


CD 


<D 


D 


3 


3 


3 


c 


c 






u. 


U. 


U. 


Li. 


LL 


LL 


UL 


U. 


o 


O 


O 


O 


CO 


CO 


CO 


CO o 




































T3 




























0) 








c 

CD 














1 

Q. 


CD 
T3 


















CD 


CD 
P 

O 

x: 
O 


c 








1 

Q. 

C 

o 

I— 




C) 


O 
Q. 

c- 


o 

o 


o 




O 


o 


CD 
Q. 

n 






o 


E 






^ 


o 

in 


o 


o 


o 


CvJ 






>* 


o 

in 


o 






o 


CO 

O 

c 
c 

CD 


Q> 
Q. 
K 




<0 

c 


o 

C 


OD 


o 


o 


CO 


CO 
"fl) 


b 

-C 
CO 


CM 


00 
CO 


CJ 

1- 


O 




O 



in 


CO 

3 
O 


111 




T3 
CD 
Q) 

CC 

U 


P 


CI) ' 


71 u 
T> C 


n 


CO 

in 

CD 

> 

CD 
CD 
CO 


(/3 

in 
> 

CD 

s 

CO 





o 


b t 


n 

D 


eg 


00 


c 


CO 




c 

OC 

c 


1 

5= 


1 

_l 


■D 

Q. 

c 
o 


Q. 
O 

o 

c 

CD 


Q. 
O 

o 

c 


Q. 
Q. 
O 
O 

§ 
Q 


CO 

a3 

Q. 
D. 

o 
O 


CO 

k- 

Q) 
Q. 
CS. 
O 

o 


Q. 
O 

A 
o 

c 

0) 


CD 

b 

x: 

CO 

c 
f- 


CD 

b 

x: 

CO 

c 

H 


CD 

to 

k_ 
Q) 

CD 
Q) 
CD 


Q. 
O 

o 

CD (- 



52 



tz 



53 
different conditions. The initial investigative 
experiments used ''tc. Technetiuin-95m was used in latter 
experiments where necessary (as explained in Section 3.1). 

To interpret this table, the third copper shot 
experiment (Copper Shot at 85°C) is described. Copper shot 
(41.89 g) was loaded into the reaction column. The reaction 
column was placed in the FT-4C, indicated by the "a" in the 
System Used column. The low pressure was set to 197 psi, 
and the high pressure was set to 249 psi. The 1.5 x lo^ ppb 

Tc-spiked J-13 solution was placed in the supply 
reservoir. The flowrate was initially set to 50 mL/day, and 
later the flow was stopped, indicated by mL/day. The 
pressure was turned on. The temperature controllers were 
set to have the column heat up to 85°C, and the heat was 
turned on. The pump was turned on. 

The solutions used for these experiments were either 
distilled, deionized water (DIW) , or simulated J-13 well 
water (J-13) . Effluent samples were periodically collected 
and analyzed to determine the technetium concentration. 

Two experiments, at a minimum, were designed for each 
metal reactant, one at room temperature and one at an 
elevated temperature. The different temperatures were 
chosen for two reasons: (i) to determine if reaction 
kinetics increased because of an increase in temperature and 
(2) because technetium migration away from a geologic 
repository, as a result of water infiltration into a 



vys^-' v,c:h . 54 

breached waste package, can potentially occur at any 
temperature below the boiling point of water (95°C at Yucca 
Mountain) . Besides changes in temperature, reaction 
kinetics were also investigated by changes in flowrates. 

3.4.1 FT-4C Flow-Through System 

The FT-4C was chosen for the initial investigation of 
the metal reactants ' redox effectiveness on technetium. It 
has all the necessary mechanics (as explained below) to 
carry out the experiments: variable flowrate settings, 
variable temperature setting, capability of safely handling 
flow-through experiments using radioactive liquids, and 
capability of taking samples of the column effluent safely. 

The equipment description for the FT-4C is presented in 
Section 3.4.1.1, the FT-4C readiness experiment is presented 
in Section 3.4.1.2, and the experiments' description using 
this system is presented in Section 3.4.1.3. 

3.4.1.1 Equipment description 

The Cortest Systems Incorporated FT-4C Flow-through 
System, Figure 11a, is a one-of-a-kind high-temperature, 
high-pressure autoclave designed for investigating 
liquid/solid reactions at flowrates from 1 mL/day to 
100 mL/day. ^^ It is shielded to allow use of radioactive 
liquids or solids, the temperature can be varied from 



- ■ ■ ■:■. 55 

ambient to SSCC, and the pressure can be varied from 100 
psi to 5000 psi.^^ 

The FT-4C operates by pumping fluid from the supply 
reservoir by a two-piston metering pump through a 
6-mm-inside-diameter by 3 0-cm-long horizontal titanium 
reaction column into the waste-collection vessel. The liner 
of the column is equipped with a 10-/im frit at the outlet 
side of the column. The system has three thermocouples that 
measure the temperature at the front, middle, and rear of 
the furnace. A bypass valve is located between the outlet 
of the reaction column and the waste vessel to allow 
effluent samples to be collected without subjecting the rest 
of the system to ambient conditions. 

The FT-4C was modified such that all tubing and 
components downstream from the pump could be flushed with an 
oxidizing agent without contacting the pump's o-rings. This 
modification was necessary because the pump's o-rings 
deteriorated with some of the oxidizing flush solutions 
used, such as nitric acid. There is a flush trap on the 
vacuum line for collecting the flush-solution samples, which 
can then be analyzed to determine the amount of technetium 
deposition along the system's tubing and components. A more 
detailed schematic diagram of the system is shown in 
Figure 13. 

The FT-4C was used to conduct the Blank-run, Copper 
Shot, Tin Shot, and four Iron Powder experiments. Solutions 



56 




o 



a> 

x: 
o 

•o 



Ql 

E 
CO 

CO 



O) 



0) 
0) 
CO 

E 
B 

U) 

(0 
o> 

O 



o 






(O 
0) 

■e 
o 
O 
(1) 

o 

E 

2 
o> 

(0 

b 

o 

CO 

E 



T3 

Q 

CO 

3 
O) 



57 
with known '^Tc concentrations were pumped through reaction 
columns in the FT-4C. Effluent samples were collected via 
the bypass valve after the solution passed through the 
reaction column. These samples were then analyzed for 
changes in technetium concentration from the inlet to the 
outlet of the reaction column. The Blank-run, Copper Shot, 
Tin Shot, and four Iron Powder experiments conducted using 
this system are discussed in Section 3.4.1.3. The 
discussions of the results are presented in Sections 4.2, 
4.4, 4.6, and 4.3.3, respectively. 

3.4.1.2 Equipment readiness experiment 

An equipment readiness experiment was conducted before 
the technetium experiments using the FT-4C to determine if 
the system's performance was satisfactory from both safety 
and operational standpoints for the radioactive experiments. 
This experiment used uranium as a surrogate for '^Tc since 
uranium solubility, like technetium, is sensitive to 
solution redox conditions. By using uranium, the system 
could be evaluated using a solution that was only slightly 
radioactive. 

A solution containing 100 ng/mL (ppb) of uranium in 
J-13 was pumped at approximately 50 mL/day through a column 
containing iron powder at 25 °C. The uranium concentration 
in the column effluent initially dropped more than 95% 
rather rapidly. it then fell less rapidly to levels below 



58 




8 

(N 



s 



? 



s 







a> 












^ 


















Q) 


^ 




CO 






1 

-5 




'w" 


■D 
0) 




il: 


Ji 


o 


3 

o 


Q. 




0) 

E 


E 

3 

'E 




H 


5 

3 
< 


s 




E 

3 

'c 
5 


o 




(A 






eriment U 
n at 25°C. 


? 




iness Exp 
ion colum 


o 




Read 
react 


CM 




FT-4C 
powder 



V t. 



(Hlu/Bu) uo!}bj}U80uoo Lunmejn 



3 






yh ai J- 



^%-H 59 



the detection limit (a 0.1 ng/mL) of the Scintrex UA-3 
Uranium Analyzer, as shown in the data in Figure 14. The 
system's performance was satisfactory from both safety and 
operational standpoints; it was also judged ready for 
experiments using technetium since the reaction reduced the 
uranium concentrations as expected. 

3.4.1.3 Experiments' descriptions 

Ten of the experiments listed in Table 4 used the FT-4C 
System (Figure 11a) . After the FT-4C Readiness Experiment, 
three copper shot, two tin shot, three iron powder, and one 
blank run experiment were conducted. * 

For all of the technetium-iron powder experiments, the 
iron powder was washed with ethyl alcohol to remove the fine 
particulates so that they would not pass through or clog the 
frit at the downstream end of the reaction column. To 
accomplish this, the iron powder was placed in a beaker 
(either 100 mL or 250 mL) , and ethyl alcohol was added to 
sufficiently cover the powder. The mixture was agitated by 
shaking the beaker. The mixture was allowed to settle for a 
few seconds before the fine particlulates suspended in 
solution were decanted off. This was repeated until the 
solution was not cloudy when agitated. The final solution 
was then poured out of the beaker, and the iron was allowed 
to air dry before loading the column. 



60 
For each of the experiments, except the Blank-run 
Experiment, the following were conducted: 

• The 6-mm-ID x 3 0-cm-long titanium reaction column was 
packed with the reactant being studied. 

• The low pressure was set at or above 179 psi. This was 
the minimum pressure necessary to keep the system 
functioning smoothly for these experiments. The high 
pressure at the inlet end of the reaction column was 
set at or above the values needed to force the liquid 
through the reactant at the chosen flowrate. 

• The technetium-spiked solution was prepared and placed 
in the supply reservoir. A sample was collected from 
the supply reservoir. 

• The desired flowrate was set. 

• The pressure was turned on. 

• The temperature controllers were set and turned on for 
the elevated temperature experiments. 

• The pump was turned on. 

• Analytical samples were collected periodically. 

The Blank-run Experiment determined whether ''tc was 
being retained in the plumbing downstream from the reaction 
column. To do this, non-spiked J-13 solution was passed 
through the lines downstream from the reaction column only. 
Samples of this solution were periodically collected to be 
analyzed for ''Tc. Following the non-spiked J-13 solution, 
0.1 M nitric acid was passed through the lines downstream 



61 
from the reaction column. J-13 was again flushed through 
the lines downstream from the reaction column following the 
nitric acid flush. Samples of this second J-13 flush were 
collected and analyzed for "Tc concentration. 

The FT-4C was flushed with 0.1 M nitric acid, J-13, 
and/or deionized water before conducting all experiments. 
For the first three experiments, the system was flushed with 
DIW as good laboratory practice. DIW was also pumped 
through the system when system maintenance was being 
conducted. After the Blank-run experiment, J-13 and/or 
nitric acid was used to flush the system. Before the iron 
powder experiments, the system was flushed with nitric acid 
and J-13 until the technetium concentration of the samples 
from the flush solution was less than approximately 
1 X 10"'' ppb. 

The results obtained from the Blank-run, Copper Shot, 
Tin Shot, and four Iron Powder experiments that used this 
system are discussed in Sections 4.2, 4.4, 4.6, and 4.3.3, 
respectively. 

3.4.2 Bench-Top Plow-Through System 

Because of some problems with the FT-4C (as will be 
discussed in Section 4.2), the Bench-top Flow-through System 
(Figure lib) was used in some experiments. The system's 
equipment is discussed in Section 3.4.2.1. The Bench-top 
System was used to conduct the copper powder, two iron 



. . 62 

powder, and the stannous chloride and tuff experiments. The 
experimental details are shown in Table 4, the experimental 
descriptions are discussed in Section 3.4.2.2, and the 
results are presented in Sections 4.4.1, 4.5.2, 4.3.1, 
respectively. 

3.4.2.1 Equipment description 

The Bench-top Flow-through System pumped a technetium- 
spiked solution through a reaction column. Two types of 
reaction columns were used in this system. One was 
identical to that of the FT-4C, a 6-mm-ID x 3 0-cm-long 
titanium column with a 10-/xm frit at the outlet end. The 
other was a stainless steel reaction column approximately 
1-cm-ID X 6-cm-long with a 2-Mm frit at the outlet end. The 
solution was pumped using a Milton Roy metering pump, which 
is a reciprocating plunger, positive displacement pump 
capable of flowrates from 50-500 mL/day. The flowrates for 
the experiments using this system were between 100 mL/day 
and 160 mL/day. All the tubing for this system was 1/16 
inch stainless steel. This system was not equipped with a 
furnace and was used at ambient conditions. 

3.4.2.2 Bench-top experiments 

For the Bench-top experiments listed in Table 4, the 
reaction column was loaded with the reactant being 
investigated. The copper powder and iron powder were washed 



63 
with ethyl alcohol as described in Section 3.4.1.3. The 

99 

Tc Iron Powder Experiment used the l-cm-ID x e-cm-long 
stainless steel reaction column; all the others used the 
6-mm-ID X 3 0-cm-long titanium column. 

Except for the Stannous Chloride and Tuff Experiment, 
the pump was initially connected to the bottom of the 
vertical reaction column for all the '5"™Tc experiments. The 
solution was pumped from the bottom up through the reaction 
column until the solution appeared exiting the top of the 
column. This ensured that it was filled with liquid, and 
the solution was not channeling through the column. For the 
remainder of the experiment, the pump was connected to the 
top of the reaction column to force the solution to flow 
from the top to the bottom of the column. Then any 
precipitate formed could be expelled from the column with 
the flow rather than collecting on the bottom of the column. 
For the Stannous Chloride and Tuff Experiment, the inlet 
solution entered the top of the reaction column only. 
Otherwise, the stannous chloride would have dissolved before 
changing the direction of the flow from the bottom of the 
reaction column to the top. Analytic samples were collected 
periodically from the outlet of the reaction column. 

3.4.3 Beaker System 

The Beaker No-flow System was used for investigating 
stannous chloride. A schematic of this system is shown in 



64 
Figure lie. The Beaker No-flow System has the reactant in 
semi-static contact with the technetium-spiked solution. 
The equipment is discussed in Section 3.4.3.1, and the 
experiment is discussed in Section 3.4.3.2. 

3.4.3.1 Equipment description 

The Beaker No-flow System, Figure lie, was used at 
ambient conditions to study the effect of dissolved stannous 
chloride on '^Tc reduction. To use this system, a 
technetium solution was added to the beaker, which contained 
dissolved stannous chloride. 

The Beaker System was used because stannous chloride is 
a soluble salt; i.e., it dissolves when in contact with the 
solution used in the experiments. By using this system, the 
dissolved stannous chloride's redox effectiveness could be 
studied because the dissolved salt would remain in the 
system rather than being pushed through a reaction column. 

3.4.3.2 Experiment description 

The Beaker Stannous Chloride Experiment used the beaker 
setup. Figure lie. Stannous chloride (10.00 g) was placed 
in a beaker containing 3 mL of J-13 and stirred vigorously 
with a magnetic stir-bar for 30 min until the stannous 
chloride was fully dissolved. Ninety-three mL of a 26 ppb 

99m 

Tc-spiked J-13 solution were then mixed into the beaker 



r-r-' 






65 
containing the dissolved stannous chloride and stirred with 
the magnetic stir-bar for 15 min. 

The first analytical sample collected was from the 
'^Tc-spiked solution before it was added to the stannous 
chloride. The second analytical sample was from the 
stannous chloride/^'Tc-spiked J-13 solution. It was 
collected using a syringe and analyzed for technetium 
concentration. A third sample was collected 24 hours later, 
placed in a filter (20 A pore size) and centrifuged until 
the solution passed through the filter before analyzing for 
technetium concentration. To ascertain that the technetium 
was not absorbed by the filter in the above samples, the 
filter for the fourth sample was primed. It was primed by 
passing a portion of the solution through the filter and 
then discarding the solution. This allowed the filter to be 
saturated with the sample. After priming the filter, the 
remainder of the solution from the fourth sample was then 
centrifuged in this same filter. The centrifuged solution 
was then analyzed for ''tc concentration to determine if all 
the technetium precipitated or if some remained in solution. 
The centrifuge tube was then washed with 0.1 M nitric acid, 
and the nitric acid solution from the centrifuge tube was 
analyzed for technetium concentration. 



'V. ■■.*- ^ f ' V T ~. 'i 



i .1 » 



3.4.4 Sealed Vessel System 



A-- rHP. ^^ 



The Sealed Vessel System, Figure lid, was used for the 
remainder of the experiments with iron powder. This system 
was designed to obtain more information on the reaction 
kinetics of iron and on the solution environment of the 
experiment. This system is described in Section 3.4.4.1, 
the readiness experiments are described in Section 3.4.4.2, 
and the experiments conducted using this system are 
described in Section 3.4.4.3. 

3.4.4.1 Equipment description 

The Sealed Vessel System used no-flow (steady-state) 
conditions. It was designed and used to determine if the 
much longer contact time between the solution and the iron 
powder, compared to the flow-through systems, would cause 
the '^'"Tc concentration in solution to drop BDL. This system 
was also used to study the variations in the solution 
environment (pH and Eh), caused by iron powder. The Sealed 
Vessel System allowed for comparison of the technetium 
concentration and the solution environment variables with 
those obtained from the EQ3/6 simulation results. 

The Sealed Vessel System consisted of a sealable vessel 
inserted into a heating mantle. The cover for the vessel 
had port holes that held rubber stoppers. The port holes 
and the vessel cover were all sealed with silicone grease 
(Dow Corning Compound ill Valve Lubricant and Sealant 



67 

containing Silicon and rated -40°C to 204 °C 

[-40°F to 400°F]) to limit the amount of air infiltrating 

into the system. A pH probe, a thermocouple, a platinum 

electrode, a saturated calomel electrode (SCE) , and a sample 

port were inserted through rubber stoppers in the port 

holes. 

The sample port consisted of a stainless steel tube 
with tygon tubing slipped over it. A clamp on the tygon 
tubing prevented the solution from leaking and also 
prevented air from entering the system. The stainless steel 
tubing was below the solution level, but above the iron 
powder on the bottom of the vessel. A pipet tip was 
inserted into the tygon tubing, the clamp was removed, 2 mL 
of solution was withdrawn, the clamp was replaced, and the 
solution was discarded. A new pipet tip was inserted, the 
clamp was removed, and a 5-mL sample was retrieved for 
technetium-concentration analysis. The tygon tubing was 
then clamped again. The first 2 mL were discarded because 
the sample port held 2 mL, and it may have contained 
solution from the previous sample. 

Two iron powder experiments used this system, one at 
7000, and the other at ambient temperature. The experiments 
are explained in Section 3.4.4.3, and the results are 
presented in Sections 4.3.2.1 and 4.3.2.2, respectively. 



■ ' 68 

3.4.4.2 Readiness experiments 

Three experiments using the Sealed Vessel System 
determined whether (1) the lack of oxygen limits the amount 
of aqueous iron such that Eh measurements were unobtainable, 
(2) meaningful Eh measurements were obtainable, and (3) the 
system was safe for conducting experiments using radioactive 
solutions. To measure the system Eh, the iron concentration 
in solution must be high enough. Eh measurements with iron 
concentrations of lO"*^ and 10"'' M have been successfully 
measured using a platinum electrode under anoxic 
conditions." Because the EQ3/6 simulations calculated 
some of the iron concentrations to be within this range, 
meaningful Eh measurements should be obtained. The supply 
of aqueous iron should not be limited because the Sealed 
Vessel System was not purged with an inert gas and was not a 
true anoxic environment. The iron should have reacted with 
the oxygen in the system resulting in obtainable Eh 
measurements. The lack of oxygen limiting the supply of 
aqueous iron to the solution is addressed further when 
analyzing the results of the technetium experiments using 
this system (see Section 4.3.2). 

The Eh of the system was obtained by using a platinum 
electrode with an SCE reference electrode, as recommended in 
the literature. 22,33 The accuracy of the response of this 
electrode combination was checked using solutions containing 
quinnhydrone. Figure 15 Curve 1 shows that the measured 



i ■M^.x,..- ..;•■■■. 

potentials were in very close agreement with the theoretical 
values, Figure 15 Curve 2.^^ It was therefore determined 
that the Eh measurements were obtained and had thermodynamic 
significance. 

During these readiness experiments, the platinum 
electrode's surface spontaneously oxidized with atmospheric 
oxygen. The platinum surface was roughed with very fine 
emery cloth to remove the oxide layer as recommended in the 
literature.-^^ 

After correcting a couple of hardware problems, the 
system appeared to function correctly and was deemed ready 
for experiments with technetium from both a safety and 
operational standpoint. 

One limitation could not be addressed before performing 
the Sealed Vessel experiments using technetium. Guppy and 
Atkinson^^ j^^^^ found that solid technetium compounds that 
precipitate or are retained on iron would prevent the iron 
from further redox reactions. This limitation was 
considered when analyzing the results of the Sealed Vessel 
experiments (see Section 4.3.2). 

3.4.4.3 Experiment description 

Two experiments were conducted using the Sealed Vessel 
System, Figure lid. The sealable vessel was first placed in 
the heating mantle. The pH probe, SCE, and sample port were 
placed into the sealable vessel's top. The platinum 



70 




I 

Q. 



<9 



«P 



c 
o 

■o 

C 

c 

o 

c 
'w 

D 

c 
g 

75 
O 

c 
,g 

re 

c 

0) 

E 

w 

c 



UJ 

E 

w 

(0 

0) 
M 
M 

> 

© 

I 



(ALU) 43 



10 



I 



71 

electrode was wiped with a #600 emery cloth and put into the 
vessel's top. Both of the Sealed Vessel experiments had a 
2.1 X 10"5 ppb 95"'Tc-spiked J-13 solution placed in the 
sealable vessel to which 32 g of iron powder were added. 
The vessel cover was placed on the vessel. The stoppers, pH 
probe, thermocouple, platinum electrode, SCE, and sample 
port were all sealed with silicone grease. The vessel cover 
was then sealed with silicone grease. For the Sealed Vessel 
Experiment at 70°C, the temperature controller was set to 
maintain a temperature of 70 °C and then turned on. The 
stopper containing the thermocouple was loosened to relieve 
pressure while the system was heated, and then it was sealed 
in place again. Samples were periodically collected via the 
sample port. 

Thirty-five g of sodium chloride crystals were added to 
the system 39 days into the Sealed Vessel Experiment at 
24 °C. At this time, the temperature controller was set and 
turned on to raise the temperature to 70 °C. The sodium 
chloride was added in an attempt to increase the corrosion 
rate of iron. Eleven days later, another 30 g of sodium 
chloride crystals were added to the system. 

For a discussion of the results obtained for the Sealed 
Vessel Experiments at 70 °C and 24 °C, see Sections 4.3.2.1 
and 4.3.2.2. .. 



py;. *■■'" *■ CHAPTER 4 

" RESULTS AND DISCUSSION 

' '* 4.1 Overview 

The results of the experiments presented in Section 3.4 
are arranged by the reactant used rather than by the system 
used to allow for continuity in the discussion of each 
reactant. Table 5 indicates the sections where the 
preparation, description, and discussion for the experiments 
are located. 

The results of the experiments are shown in 
Figures 16-42. These figures show the times at which 
flowrates were changed, when applicable, when the technetium 
concentration went BDL, and when the NaCl was added to the 
Sealed Vessel Experiment at 24 °C. Some figures show 
comparisons with the results obtained from the EQ3/6 
simulations. It may be necessary to review Section 2.3.2 
for interpretation of the figures that include the EQ3/6 
simulations. 

The experimental results are also summarized in 
Table 6. The table is grouped by reactant and shows the 
initial, final, highest, and lowest technetium 
concentrations as well as the time these concentrations were 
obtained and the total time the experiments were conducted. 



72 



;. ^ ■"i 



C 
o 

•H 
■P 

ns 
O 
O 

h^ 

■P 

C 
0) 

•H 

0) 

a 



ID 

0) 

rH 

XI 

to 





n 0) 
































s-d 
































-H 










H 


(M 




H 


oo 














a M 










• 


• 




f 


• 














a 










(M 


CN 


fNl 


fO 


ro 


H 












n> H 










• 


• 


• 


• 


• 


. 












■P 43 










■* 


•^ 


IT) 


•^ 


"^ 


ID 












CO u 










r) 


n 


•* 


n 


n 


'4' 
















H 


n 


























c 


(4 


• 


• 


























•H 


« 


H 


iH 


























H 


E 

c 




• 


• 

• 


























































•H 
































■p 






























0) 


u 

0) 
CO 


• 


• 

H 


CM 


H 


(M 
CM 


H 




















o 




• 


• 

n 


• 


« 


• 

n 


'1' 

• 

-3" 






















H 


n 




H 


CM 










H 


n 













• 


• 




• 


• 












• 













H 


H 


n 


(N 


(N 


H 








^ 


^ 


CNJ 






M 




• 


• 


• 


• 


• 


• 








• 


• 


• 






H 




^r 


'* 


n 


■^ 


•^ 


n 








«* 


■^ 


n 










n 


n 


"a* 


ro 


n 


^ 








n 


n 


'3' 






a 




C 


C 




C 


C 




C 


C 




C 


C 













O 





C 








C 


O 





c 





O 


c 




•H 




•H 


•H 





■H 


•rA 





•H 


•rH 





•H 


•rH 


n 




0) 




■P 


-P 


•H 


+J 


-P 


•H 


■P 


-P 


•H 


+J 


+J 


•H 




0) 




(0 


a 


0) 


(0 


a 


w 


(0 


ft 


tfl 


(fl 


Oi 


fO 




0) 




v^ 


•H 


w 


u 


•H 


tn 


^H 


•H 


[fl 


u 


•H 


fn 




M 




(C 


V^ 


d 


(0 


^ 


3 


(fl 


^H 


D 


(Tl 


^ 


T 




&< 




ti, 


u 


u 


a 





U 


a 





() 


ft 


D 


C) 









(1) 


UJ 


w 


Q) 


U) 


Ul 


Q) 


U) 


t/) 


0) 


m 


rn 




M 




^ 


<u 


•rH 


Vh 


Q) 


•H 


U 


a) 


•H 


^H 


0) 


•H 




fu 




tu 


Q 


a 


a, 


Q 


Q 


CU 


Q 


Q 


0^ 


Q 


Q 
























e 






rH 


























<u 






0) 








0) 


















-p 






rn 








+» 


















t/i 






rn 








c 












a 






>i 






0) 








0) 



















CO 






> 








6 












-p 




















•rl 












1 






u 






T( 








h 






u 






sn 






Q) 






0) 








V 






'* 






u 






^ 






rH 








ft 






1 






c 






fO 






(0 






K 






H 






Q) 






0) 






(1) 






H 






U< 






pa 






OQ 






CO 




» 



73 



-p 

0) 

■p 

(0 

tr« 

•H 
-P 

CO 

0) 

> 

c 

•rH 

o 
-p 

T3 
Q) 
CO 

;3 

-P 
O 
C 

m 

(0 

e 

Q) 
■P 
W 
>i 
(0 

•H 

-P 

Ui 
Q) 
■P 

(0 

u 

•H 
Tl 
C 

•H 

^ 

(U 
X! 
E 

3 
C 

C 

O 

•H 4-J 

-P C 

O (0 

Q) -P 

cn o 
o 



+> 

CO 
0) 



(0 

c 
g 

•H 

a 
w 



VD 
(U 
X! 
Eh 



Total 
Time 
(days) 


CO 
CM 


o 
CO 


in 


CO 
CM 


CO 


CO 


o 


in 


CD 

in 


00 


03 


in 


CO 


t^ 


CD 
CM 


- 


CM 


h- 




Time of Lowest 
Concentration 


CO 

> 

CO 

•o 

CO 
CM 


CO 

> > 
to 

T3 


CO 

- > 

CO 

■o 

CM 


CO CO 

>> > 

CD CO 

-o -a 

CO ^ 
CM CO 


CO 
> > 

CO 
T3 

q 


CO 
> 

CO 

•o 
o 


c 
E 

o 


CO 

> 

CO 

■o 


£1 
CM 


O 

d 


CO CO 

>< > 

CD CD 
T3 -D 

N. o 

r^ CO 


sz 

o 
CO 


in 


3.0 days 
1 .0 days 
5.4 h 




Lowest Tc 

Concentration 

(ppb) 


_l 

Q 

CQ 


o 

X 

00 


o 

X 

in 


o 

X 

CM 


o 
X 

CO 

in 


_I 
Q 
CO 


o 
X 

CM 


_I 
Q 
CO 


_l 

Q 

CO 


o 
O 

X 

o 

CD 


o 
X 

CM 


O 

X 


o 
X 


9 
O 

X 

in 


CO 
CM 


CM 

O 

X 

CD 
00 


_I 
Q 
CO 


o 

X 

CO 

CD 




Final Tc 

Concentration 

(ppb) 


Q 
CO 


o 
X 


o 


o 
X 


00 

lO 


_I 
Q 
CQ 


in 

o 
X 

CO 


_J 
Q 
CO 


_I 
Q 
CO 


n 

o 
X 

CM 


o 
X 

CvJ 


m 

o 

X 

CO 


o 
X 


o 
X 


CM 

o 
X 


o 

X 

q 


_l 
Q 
CQ 


en 

o 
X 

CM 


Q} 

in 
cu 
> 


Time of 

Highest 

Concentration 


o 


CO 
CO 

■o 

o 
CO 

CM 


o 
d 


o 
d 


o 
d 


o 
d 


o 
d 


o 
d 


o 
d 


CO 

■o 
CO 


CD 
T3 

CO 


CO 

>^ 

CD 
•D 

CO 


CO 
CO 


CO 

>s 

CO 
T3 

O 


o 

d 


CO 
T3 

in 


o 
d 


CO 

>. 

CD 
T3 

O 


T3 

0) 
CO 
OJ 

to 


Highest Tc 

Concentration 

(ppb) 


o 
X 




o 

X 

in 


o 

X 

CO 
en 


o 

r— 

X 

in 


o 

X 

CD 
CO 


'J 

o 

X 


in 
O 

X 

CO 
CM 


up 

o 

X 

CO 

CM 


n 

o 
X 

CO 


o 

X 

"3; 


n 

o 

X 

in 


o 

X 

o 

CM 


o 

X 

q 


o 

X 

q 


o 

X 

q 


o 

X 

CD 
CM 


o 
X 

CM 


CO 
CL> 
CO 


Initial Tc 

Concentration 

(ppb) 


o 


o 


o 

X 

in 


CM 

o 

X 

CD 

ai 


o 

X 

in 


O 

X 

to 

CO 


O 
X 

d 


in 

o 
X 

CO 

CM 


in 

o 
X 

CO 
CM 


o 
X 

CM 


o 
X 

CM 


m 

o 

X 

CO 


r-i 

o 

X 

in 


io 
X 


o 

X 

q 


o 

X 

q 


O 

X 

CD 

CM 


■r 
o 

X 


O 

x: 
u 

c 

OJ 

CD 
J3 


System 
used ^ 


(0 


(0 


(0 


CO 


CO 


J3 


n 


■a 


•D 


CD 


CD 


CO 


CD 


ID 


CD 


CO 


o 


XI 


t 

to 

in 

E 

QJ 
Ul 
>, 

(/) 

-2 

c 

£ 

a 

X 

UJ 


Experiment 


CO 

c 
ca 

to 

00 


•o 
o 

< 

o 

z 

c 
a: 

c 
CO 

CO 


1 

Q. 

C 

2 

§ 

o 

I 


Q) 

■a 
1 

CL 

C 

o 

i— 

1 
§ 


O 

o 

in 

CO 
CO 

k_ 

CD 

■o 

1 
Q. 

c 
o 


CD 

1 
O 
Q. 

c 
o 

o 

1- 

n u 
n c 


CD 

■o 

Q. 

c 

o 

k_ 

o 

1- 

1 

n 


O 



O 
CO 

L_ 

CD 
T3 

1 

a. 

c 
o 

1— 


O 
o 

CM 

CD 

k_ 
CD 

1 
O 
Q. 

C 

2 


O 

s: 

CO 

k_ 
CD 
D. 
Q. 
O 

o 

Q 


_1 

E 

Al 

O 
JC 
CO 
k_ 
CD 
Ql 

a 
o 
O 

Q 


O 
o 

';f 
CM 

TO 

O 
JZ 
CO 
k- 
OJ 
Q. 
Q. 
O 
U 


O 

o 

in 
00 

to 

o 

CO 

k_ 
CD 
CI 
Q. 
O 

o ^ 


0) 

1 

Q. 

k_ 
CD 
CL 
Q. 
O 
O 

o 

H- 


O 
o 

CM 

Id 

sz 

CO 

c 
1- 


O 

o 

in 

00 
CD 

o 

CO 

c 

h- 


CD 

.■g 
o 

o 

CO 

o 

C 

c 
iS 

CO 


Stannous Chloride and 
Tuff 



74 



75 
It is evident from Table 6 that none of the experiments 
that used the FT-4C system had the technetium concentration 
fall to BDL. It was concluded from the Blank-run Experiment 
(see below) that the FT-4C retained technetium from previous 
experiments and then released some of the technetium in 
subsequent runs. 

Iron was the only reactant to reduce the ''tc 
concentration BDL in the flow-through experiments. In the 
no-flow experiments, stannous chloride reduced the ''tc 
concentration BDL («0.4 ppb) , and iron reduced the '^""Tc 
concentration BDL(ssl x lo'^ PPb) . 

Experimental details and discussions are contained in 
the following sections. 

4.2 Blank-run Experiment 

To clarify the discussion of this Blank-run Experiment, 
a chronological listing of the experiments using the FT-4C 
is shown in Table 7. 

It was suspected that the FT-4C, which was shown in 
Figure 11a, was retaining technetium and releasing it later 
in the same or subsequent experiments. Experiment 3, the 
Copper Shot Experiment at 85 °C (which will be discussed in 
Section 4.4.2.3) had an initial increase in technetium 
concentration. However, the "tc concentration analysis for 
this experiment was not completed until a few days before 
Experiment 4, the Tin Shot Experiment at 24 °C. The flush 



m 
-p 

c 

u 

w 
u 

Eh 

O 

>i 
Cn 
O 

r-i 

o 

c 
o 

u 



(0 
Eh 





Initial 

Concentration 

(ppb) 


CM 

o 
X 




o 
X 




o 

X 

CO 






n 

o 

X 

in 




en 

o 

X 

q 




■ 


o 

X 

in 




o 

X 

in 




o 

X 

p 




o 

X 

in 




o 

3 
C 


to 


CM 




h- 




o 






o 

1- 

cn 
en 




o 




I 


o 

1- 

CT) 




o 




^ 

S 




o 

1- 




0) ^ 

- E 


CO 
00 




o 

6 




CD 

6 
IT) 






CD 
CD 

in 




o 
in 




o 
in 


o 
o 

6 
in 

6 




o 
in 




o 
in 




o 
in 

6 




3 

E 


in 

eg 




■<1- 

C\i 




CM 






in 
00 




CNJ 




in 

CM 


CM 




in 
00 




s 




CM 




c 
cc 


o 

CO 




s 
§ 




CO 

5 






oo 

5 




in 

CD 

in 

CO 




■ 






d 
CO 




CD 

CD 
CO 




CD 

ai 

C\J 




n 
CC 


Q) 

1 

a. 

c 
o 


i 

E 
to 

Q 


o 

x: 

Q. 
O 
O 


CO 

z 

X 

CO 

Q 


o 

CO 

k. 
0) 
Q. 
Q. 
O 

o 


Q 


o 
5 

Q. 

E 

(0 

3 
!*: 

D 


o 

CO 

k_ 

(U 
Q. 
Q. 
O 

o 


CO 

3 
*♦— 

Q 


o 

CO 

c 
i- 


■D 

CO 

E 

3 
3 
O 

E 
w 

C 

5 

o 
n 

O 

z 


O 

z 
I 

+ 

CO 


0) 

I 

c 
o 


CO 

+ 

o 

z 
I 


Q. 

C 

2 


CO 
0) 

if 

5 o 

« s 

<D 00 

|i 


o 
x: 
to 

c 

i- 


CO 

3 

•D 
O 

to 

CO 

+ 

o 

z 


CD 

1 
a 

c 
o 




c 


0) 

a 

>< 


CO 

« 
0) 

c 

(D 
Q) 
DC 

O 

a! 




o 

s: 

Q. 
Q. 
O 

o 

Q 




O 



CM 
CO 
O 

CO 

k. 
<D 
Q. 
D. 
O 

o 






o 



S 

o 

CO 

Q. 
O 

O 




O 



•V 

C\J 

to 

O 

CO 

c 
i- 




c 

3 
CC 

CO 

oo 


1 

Q. 

c 
o 

o 

i 




o 

o 

in 

CO 

(5 

1 

CL 

c 

2 




O 

o 

in 

00 

ra 

o 

x: 
CO 

c 
i- 




1 

c 
o 

§ 

1 










- 




CM 






CO 




n- 




in 


CO 




1^ 




00 




c:^ 



76 



77 
solution between Experiments 3 and 4 was not analyzed until 
after Experiment 4 was started. Therefore, it was not until 
after the start of this latter experiment that the FT-4C was 
suspected to retain and release technetium. Rather than 
terminate the experiment, the Blank-run Experiment was not 
conducted until after the completion of Experiment 4. 

In the Blank-run Experiment, J-13 water was pumped 
through the system downstream of the reaction column and 10 
effluent samples were collected over a period of 23 days. 
These samples were then analyzed for '^Tc concentration. 
Since this experiment did not have technetium added to the 
J-13 solution, no ''tc should have been detected in the 
sample effluent. However, as shown in Figure 16 Curve 1, 
the first effluent sample collected had a ''tc concentration 
of 14 ppb. Within the first hour, the ''tc concentration 
decreased to 0.7 ppb. The technetium concentration in the 
effluent samples then fluctuated between 0.02 ppb and 
0.3 ppb over the next 23 days. The technetium concentration 
was BDL after flushing the system with J-13 for 23 days. 

Nitric acid (0.1 M) was then flushed through the lines 
downstream from the reaction column for 5 h to determine if 
all the technetium had been leached by the J-13 flush. No 
samples of the nitric acid flush were collected because 
colleagues suggested that the detection limit would be very 
high because the fluorescence would be absorbed by the 






( I . 



■w 



78 




CO 

E 



(qdd) uouBJjuaouoo oinjiaugosi 



e 

O) 



. 79 

nitric acid.^ Following the nitric acid flush, J-13 was 
again flushed through the lines downstream from the reaction 
column. Over the next 2 days, 2 3 samples of this second 
J-13 flush were collected and analyzed for ''Tc 
concentration. As shown in Figure 16 Curve 2, the first 
sample of this post nitric acid flush had a technetium 
concentration of 7.8 ppb, which indicated that not all the 
technetium was completely removed with the first J-13 flush. 
The technetium concentration dropped to 0.08 ppb 2 days 
later. 

This experiment confirmed that technetium was indeed 
being retained within the plumbing downstream from the 
reaction column in the FT-4C Flow-through System. A 
technetium concentration below the 0.08 ppb lowest post- 
nitric acid flush was never reached during any of the 
experiments using this system. It was presumed that this 
was due to the demonstrated capacity for the system to 
retain and then re-release technetium rather than a failure 
of the reducing agent to lower the technetium concentration. 

Because experiments were conducted before realizing 
that the FT-4C was retaining technetium, the effects of this 
retention on those experiments had to be considered when 
analyzing the results. The effects of this retention 



^This turned out not to be true, and later experiments 
analyzed samples containing nitric acid. 



•t *' M *■ * 'f 



80 
phenomenon will be discussed at length in the applicable 
sections. They are briefly stated here. ' 

The technetium retention had no significant impact on 
the DIW Copper Shot Experiment, the Copper Shot Experiment 
at 24 °C, or the Tin Shot Experiment at 24 °C (Sections 
4.4.2.1, 4.4.2.2, and 4.6, respectively) because the 
detected technetium concentrations were much greater than 
14 ppb (the highest technetium concentration detected in the 
Blank-run Experiment) . The system was also flushed before 
each experiment either as good laboratory practice or for 
extended periods of time as an integral part of system 
maintenance. However, the technetium retention affected the 
Copper Shot at 85 °C and the iron powder experiments even 
though the system was flushed with DIW before conducting the 
Copper Shot Experiment at 85 °C and with nitric acid before 
conducting the iron powder experiments. The minimum 
technetium concentrations detected in these experiments were 
close to the minimum detected (0.08 ppb) in the Blank-run 
experiment following the nitric acid flush. Therefore, the 
minimum technetium concentration in the copper shot and iron 
powder experiments may have been an artifact of the 
technetium retention and release by the system rather than a 
measure of the reducing capabilities of copper and iron on 
the technetium. The effect of the technetium being re- 
released during these experiments is discussed in Sections 
4.4.2.3 and 4.3.3, respectively. 



81 
Because the FT-4C retained the technetium, the Bench- 
top System, shown in Figure lib, was designed and also used 
for flow-though experiments. 

4.3 Iron Powder Experiments 

Three different experimental systems were used to 
investigate iron powder as a reducing agent to lower the 
concentration of technetium in J-13 water: the FT-4C, the 
Bench-top Flow- through, and the Sealed Vessel systems, 
Figure 11a, b, and d, respectively. Section 4.3.1 discusses 
the results obtained from the Bench-top Iron Powder 
experiments. Section 4.3.2 discusses the results obtained 
from the Sealed Vessel experiments. The results from the 
FT-4C experiments (Section 4.3.3) are included for 
completeness even though no direct conclusions can be drawn 
from them because the FT-4C was determined to be inadequate 
for these experiments. A summary of the iron powder results 
and conclusions are presented in Section 4.3.4. 

4.3.1 Bench-Top Iron Powder Experiments 

Two Bench-top experiments were conducted; one used 
''Tc, and the other used '^mrpc. m the first experiment, a 

99 • 

Tc-spiked solution was pumped through a reaction column 
filled with iron powder at 24 °C. As shown in Figure 17, the 
Tc concentrations dropped to near detectable limits 



;j 



82 



Q 



i 



., 



8 



(qdd) uojjBJjuaouoo Lunuautioai 



I 
2 



O 





<u 




I 




t 




w: 




0) 


w 


i: 


>s 


♦-• 


(0 


1— 


•o 


(U 




c 


q; 


CO 


E 





c 
E 



CD 

E 



I 



83 
(0.4 ppb) within 25 min and remained at or below detectable 
limits for the entire 8 days the experiment was conducted. 

This was the first experiment conducted using the 
Bench-top system after the Blank-run Experiment on the 
FT-4C. The low technetium concentration found in this 
experiment confirmed that the iron experiments using the 
FT-4C were affected by the release of "Tc retained 
downstream of the pump. 

Because the ''tc concentrations in the first Bench-top 
experiment were BDL, '^""Tc was used in the next Bench-top 
experiment to allow for the 8 orders of magnitude 
improvement in the detection limit. The concentrations of 
^""Tc (ppb) and iron (ppm) for this experiment are shown in 
Figure 18, Curves 1 and 2, respectively. Curve 1 shows that 
the '^""Tc concentrations in the effluent samples rapidly 
dropped approximately 4 orders of magnitude to 1 x lo"'' ppb 
(right at detectable limits) within the first 6 h of the 
experiment and remained near 1 x lo'' ppb for the first 
13 days of the experiment. Then the concentration steadily 
increased over the next 22 days to approximately 
1 X 10"^ ppb before jumping to about 3 x lO'^ ppb for the 
last 2 days of the experiment. 

The reason for the increase in the technetium 
concentration after the first 13 days was most likely 
because the iron was sorbing the technetium and thereby 
passivating the iron surfaces, as will be discussed in more 



84 



? 




in 

CO 



o 
o O 



8 



8 



CO 

E 



K:.* ' 



(0 

c 
o 



c 
<u 
u 
c 
o 
O 
c 
o 



o 
(0 






c E _ 
.2 Q. ~ 
^ Q. E 

■5 "^ -■ 

'"Bo 



=^ 



in ' 

U (U 



3 
a! 
Q 

01 



I 



r 



1/5 

9 



9 



CO 
o 



o 



o 

UJ 



CO 
LLJ 



suo!}Bj}U8ouoo (Ludd) uoj| puB (qdd) oinjjauijoai 



o 



Q. 
O 

}- 



& 



£ 

g> 



■^' '• : s ^ - - . 85 

detail momentarily. The reason for the increase in the 
technetium concentration on day 37 is uncertain. The only 
sure means of knowing if the technetium concentration would 
have remained at this high level would have been to extend 
the experiment, but the experiment was terminated before the 
last two samples were analyzed, i.e., before the high 
technetium concentrations were known. 

A few samples were chemically analyzed for iron 
concentration to allow for comparison with the computer 
simulations. The results obtained from the experimental 
iron analyses are shown in Figure 18 Curve 2 . The first 
three samples had iron concentrations BDL (0.02 ppm) . The 
last three samples had measurable iron concentrations that 
will be discussed later along with the predictions of iron 
concentrations by the codes. . 

After the '^"'Tc Bench-top Experiment was terminated, the 
'^"™Tc activity along the length of the column was measured 
using the TCA. The results are shown in Figure 19. The 
teflon holder was 16.5 in. long, and the reaction column, 
which was only 12 in. long, was placed inside the teflon 
holder. As shown in the figure, the '^"'Tc activity was 
concentrated at the inlet end of the reaction column (left 
hand side in the figure) where it peaked at 1.4 x lo*^ Bq. 
However, there was also measurable activity (3.1 x lo"^ Bq) 
at the outlet end, which shows that technetium made it all 
the way to the end of the column. This was in agreement 



86 




(bg) AnAjpv LunjiauMoai 



a> 

ii. 



87 
with the observed increase in technetium concentration in 
the effluent samples after day 13 (as was shown in 
Figure 18) . • . , 

The high activity at the inlet end of the reaction 
column was caused by the technetium concentrating at the 
front of the reaction column as it was sorbed onto the 
active iron surface, as well as reduced and precipitated by 
the iron. The iron surfaces must have been at least 
partially passivated by sorption of technetium because the 
iron became less effective at decreasing the technetium 
concentration as time passed. This is evident by the 
increased technetium concentration in the effluent samples 
after day 13 and because of measurable technetium activity 
at the outlet end of the reaction column. The precipitation 
of the technetium between the iron particles would not be 
expected to passivate the iron. Because the iron surface 
was passivated, the new incoming technetium had to traverse 
further down the column before reaching active iron 
surfaces. Eventually, the end of the column was reached. 
At that point, technetium activity was present along the 
entire reaction column, but the activity was greater at the 
inlet end rather than the outlet end. If the reaction column 
had been analyzed daily, the first day would have shown 
technetium activity only near the inlet. Day 2 would have 
shown the technetium activity a little further down, etc. 
Eventually, it would have shown that the technetium activity 



88 
progressed all the way to the outlet end of the column. 
However, daily analyses would have required an impractical 
reconfiguration of the experimental setup to provide for low 
levels of background radiation. 

Also, some technetium was always detected in the 
effluent samples (just at detectable limits for the first 13 
days, then well above detection limits until the end of the 
experiment) because the slow reaction kinetics of iron 
prevented all the technetium from being reduced. This was 
the case even when it could react with the bare iron surface 
as it traversed down the reaction column. The slow reaction 
kinetics of iron will be apparent when the Sealed Vessel 
experiments are discussed in Section 4.3.2. 

Results from the two Bench-top Iron Powder experiments 
agree in the sense that the technetium concentrations 
dropped rapidly in both cases. In the ''tc experiment, the 
concentrations rapidly dropped below the detection limit of 
about 0.5 ppb and could have been as low as in the repeat 
experiment with '^""Tc (1 x lo"'' ppb) . 

The Bench-top '^'"Tc Iron Powder Experiment was modeled 
by the two EQ3/6 simulations, Free Run and Limited Run, that 
were shown in Figure 8 and explained in more detail in 
Section 2.3. The initial amounts of iron and technetium 
entered into the "Free Run" simulation were the same as 
those used in this Bench-top '^^'"Tc Iron Powder Experiment. 
In the other simulation, "Limited Run," the same initial 



89 
amount of technetium was entered as was used in the 
experiment, but the initial amount of iron was limited in 
the simulation such that the final calculated technetium 
concentration was equal to the value found in the 
experiment. Figure 2 shows the combination of Figure 18 
and Figure 8 for comparative purposes. 

In Figure 20, the technetium and iron concentrations 
are plotted along the axis of ordinates. Concentrations 
from the Bench-top '^'"Tc Iron Powder Experiment versus time 
(lower axis of abscissae) are represented by Figure 20 
Curves 1 and 2, respectively. Concentrations from the Free 
Run simulation versus the reaction progress (upper axis of 
abscissae) are represented by Figure 20 Curves 3 and 4, 
respectively. And finally, concentrations versus reaction 
progress from the Limited Run simulation are represented by 
Figure 2 Curves 5 and 6, respectively. 

Comparison of results from the experiment and the EQ3/6 
calculations shows that the maximum iron concentrations were 
approximately the same for the two cases, Figure 2 Curves 2 
and 4, respectively. The calculated iron concentration 
peaks at about 0.2 ppm at a reaction progress of 0.00128. 
The measured iron concentration reaches this value after 31 
days. 

Although the measured minimum technetium concentrations 
were quite low, just at detection limits of i x lo'^ ppb 
(Figure 2 Curve 1) , they did not decrease to the calculated 



90 



to 
o 
o 
d 



CM 

8 

d 



§ 



O 
O 



.H 



o 



■^ 8 



o 
o 
d 




,.□ 






□ 



fiSi 



f 



a 






a 



D 



■a-B 



II 



'" ' ■■■"■-^^-■«»— «— »»-<*P "- Q »- - ^ 



8oS8S88^gg°:r^2 

UjUJUJLUtULlJUJuJLLJuJuJliJLUUJ 



.E a: 

& 2 
o> Ul 



., 





3 






0) 


O 


ft: 


cr 


(C 


§ 










p 


3 


■^ 


f- 






n 


U> 


X 






(l> 



CO 

o 



c 
o 



« 













to 


o 






f* 


c> 




i i 


■4^ 






O 








O 






Z 






s = 








s-^ 


M 








C 






U. 1- 


o 
























CD 


in 






1 


3 












b 






■ 


CO 








CO 
















CO 








O 




(A 




UJ 




>N 


E -?- 






CD 


§■ i 


■n 


S 


3 


^^ Q. 


c 

CQ 




(U 


•2 i 


^rf 




E 


1 ir 

o aj 


C 
0) 




1- 


W S 


f- 






■E it 


•c 






a a 


s 



X >> 

HI a> 
._ > 

^ts 

I ^ 
Q. 0) 

c "". 
h- CO 

I CO 
a. (i> 

^i 

OJ to 

"^ 53 
<u > 
£ o 

v_ (O 

c c 
o CO 

CO (o 

Q. (U 

E ^ 

O 3 

O O 



o 



suojjBJiueouoQ (Ludd) uoj| pue (qdd) Lunjjaui^oai 



3 
O) 



91 
concentrations of 1 x lO'' ppb to 1 x 10'^^ ppb, (Figure 20 
Curve 3) . By the 15^^ day, the technetium concentrations 
were well above detectable limits and rising. As previously 
stated, the rise in technetium concentration in the 
experiment was most likely caused by the technetium sorbing 
onto the iron, thereby partially passivating the surface and 
causing the iron to become less reactive. The codes did not 
show this rise in technetium concentration because the codes 
do not model sorption. 

The measured technetium concentrations were also orders 
of magnitude greater than that indicated by the simulation 
for a given iron concentration in solution. According to 
the EQ3/6 results, an iron concentration of 0.064 ppm 
corresponds to a technetium concentration of 9.6 x lo"' ppb 
(Figure 20 Curves 4 and 3, respectively), at a reaction 
progress of 0.0013 (upper axis). However, in the Bench-top 
'^""Tc Iron Powder Experiment at 29 days (lower axis), the 
same iron concentration corresponded to a technetium 
concentration of 1.5 x lO"*^ ppb as shown in Figure 2 Curves 
2 and 1, respectively. 

The analysis of the reaction column (see page 85) also 
indicates that the technetium was sorbed by the iron, which 
caused the iron surface to become partially passivated and 
less effective at decreasing the technetium concentration as 
time passed. Other researchers have also reported 
technetium sorption by iron. For instance, Gu and Schulz 



92 
reported that the sorption of technetium was "associated 
with both soil organic matter and iron and aluminum 
oxides. "^^ They attributed this to "nonspecific sorption," 
which "occurs at the localized positive charges that occur 
on free hydrous oxides of iron . . . which attract anions 
electrostatically, and the anions are readily exchangeable 
with those in the soil." 

Meyer, Arnold, and Case studied the effect of basalt 
and basalt plus iron on technetium concentration. ^^ Three 
types of basalt were tested: a virgin basalt, one with iron 
powder added, and one that was contaminated by iron when the 
basalt was crushed. In the virgin basalt sample, 
significant technetium sorption did not occur. The small 
amount of technetium that was adsorbed on the virgin basalt 
was mostly present in the (VII) valence state with only a 
small fraction in lower valence states. Therefore, the 
technetium concentration was not significantly reduced nor 
was there significant reduction in the oxidation state of 
the technetium that was sorbed. However, in the basalt with 
the iron powder added, all of the technetium was essentially 
removed from the solution, and all but a few percent of the 
technetium that was retained by the basalt was in a reduced 
oxidation state. In the experiment conducted with the 
basalt that was contaminated with iron during crushing, 
essentially all of the technetium was removed from solution. 
Most of it was reported to be present as Tc(Vll), which 



.. - ■ . 93 

probably was an artifact of the analytical method used. 
Meyer et al. noted that the valence state analyses for 
technetium were not conducted under anoxic conditions. They 
attributed the Tc(VII) found on the basalt to technetium 
being easily reoxidized during its extraction. 

Meyer et al. also showed that hematite sorbed 
technetium from solution in anoxic test tube experiments.^^ 
The amount of technetium that remained in the solution was 
very small and was present as reduced technetium, but they 
do not indicate the reduced technetium's oxidation state. 
Meyer et al. concluded that a surface reaction may have been 
responsible for removing technetium from solution because 
the reduction in technetium concentration was very sensitive 
to the surface conditions. Meyer et al. also noted that if 
the oxidation state of the technetium could be reduced, then 
the technetium would be adsorbed on any surface. Again, 
this suggests that a surface reaction may be ultimately 
responsible for removing the technetium from solution. 

4.3.2 Sealed Vessel Experiments 

Although the Bench-top '^'"Tc Iron Powder Experiment 
indicated that the iron would reduce the technetium 
concentration, this experiment could not be adequately 
modeled by the codes. Therefore, the Sealed Vessel system 
was used next because the codes can model a closed system. 
The term "closed system" refers to no replenishment of the 



94 
reactant or solution (iron, j-13 plus technetium, etc.)/ and 
the solution volume remains constant. In the experiment, 
the solution volume was constant except for the removal of 
the samples, the total volume of which was relatively small. 
The pH and Eh were measured in the Sealed Vessel System for 
comparison with those calculated by the codes. To try to 
gain information on the reaction kinetics between the iron 
and technetium, experiments at two temperatures were 
conducted using this system. The Sealed Vessel Experiment 
at 70°C is presented below, and the Sealed Vessel Experiment 
at 24 °C is presented in Section 4.3.2.2. 

4.3.2.1 Sealed Vessel Experiment at 70 °C 

The Sealed Vessel System was essentially a closed 
system and was modeled as such by EQ3/6 (see Section 2.3). 
Because the experiment was started at room temperature and 
then increased to 70°C, the EQ3/6 simulation had the same 
temperature change. Figure 21 shows the technetium 
concentration. Figure 2 2 shows the system pH, and Figure 2 3 
shows the system Eh. Figure 2 4 is the combination of 
Figure 22 and Figure 23. All the figures show both 
experimental results and computer simulations. in all these 
figures, the experimental data are plotted versus time 
(lower axis of abscissa), and the calculated data are 
plotted versus reaction progress (upper axis of abscissa) . 



95 



8 

d 



i 



in 

CO 



O 



CO 

8 

d 



^^ 


lO 


N 


CN 




< ) 








d 



01 



CO 



8 

d 



s. 



w 



o 

n 
O 



f 



I 



(U 

a 

I 
m 



CO 
T3 



^ .i 



MUlllt I fcUllll I Hi..... . >. fc..., , 



CO 


■<»■ 


ID 


v 


9 


9 


UJ 


UJ 


LU 



uJ 



(seioujuj) pajejidpajd "O^oi '(qdd) uojjBjjuaouoo Lun^auijoai 



g 




3 




^ 




d) 




c 








o 




Q} 




1- 




d 




o 








CO 




3 




E 




a> 




m 




« 




£ 








? 




<D 








O 




e 




O 




r^ 




13 




4^ 




c 




0) 




E 






(- 


(U 


-1 


u. 


o 


X 


n 


LU 


F 


<1> 


o 


w 


o 




E 


> 


3 


T? 


(I) 


0) 


C 


(0 


£ 


(11 


o 


CO 


(1) 


^ 




CM 




S 




3 




.9 




u. 





96 




c 
> 

TO 
(U 

0) 
OJ 

v> 
o 

"D 
O 



T3 

0) 



o 

0) 



c 
o 



CO 



CM 
CM 

£ 

at 



97 




> 

TO 

■D 
O 



T3 

0) 



Q. 
O 



(A) 43 



-;w5i' 



98 



8 

d 



o 



o 

CO 



T3 



ra 


^^ 




CO 


o 


(1) 


§ 




"rS 


<) 










b 




W 




<n 








p) 




O 




LU 




X 




Q. 





11 

< 

■4 



o o 

S S .2 .2 



-^ '^ c 



m ^ ^ 



E E " 



E 
0) 



tU UJ lij 



uj a. ILI 



O O 



* 13 



i 




1 



<u <u 
> £ 



O 



S 

CD 
CO 



(O o 

o =- 



in 

r^ CD 

0) 

E 



CO 

V) 

tj 

3 

o 

Q. 

(C 

♦^ 

9. 
I 



E 

0) 

w 

>« 
(0 

•o 

CD 
10 
O 



o 

3 

E 
w 

CO 

O 

LU 
0) 



CO 



"? 



(A) 43 PUB Hd 



' 99 

Figure 21 Curve 1 shows that after 1 day, the 
technetium concentration fell to BDL (1 x lo"'' ppb) , which 
is in approximate agreement with the calculated minimum of 
2 X 10"^ ppb shown in Curve 2. 

Figure 22 Curve 1 shows the measured pH values from 
this Sealed Vessel Experiment at 70°C. The pH probe was 
calibrated just before the experiment using pH standards of 
4, 7, and 10. After the experiment ended, when the pH probe 
was again checked with these same standards, it gave 
readings of 3, 6, and 9. To correct for this systematic 
error, it was assumed that the pH probe had drifted 
uniformly over the 14 days of the experiment. Therefore, 
the 1 pH unit drift was divided by the number of days for 
the experiment to obtain an average drift per day. The 
corrected pH was then calculated by multiplying the drift 
per day times the day of an individual pH measurement, and 
that amount was added to the corresponding pH measurement. 
These corrected values, which are plotted in the figure, 
show the pH of the system to be between 7.9 and 9.5. These 
values correspond quite well with those predicted by the 
EQ3/6 codes, which were between 7.0 and 9.2, Figure 22 
Curve 2 . 

Figure 2 3 Curve 1 shows that the measured Eh decreased 
markedly the first day but, only slowly thereafter. The Eh 
never reached the low values indicated by the codes, 
Figure 23 Curve 2. It is believed that the measured Eh did 



not reach the low levels calculated by the codes because 
oxygen was leaking into the system. The excess oxygen 
caused the measured Eh values to be much greater than the Eh 
of the iron redox reactions. 

The suspected leakage of oxygen into this experimental 
system was discussed with Dhanpat Rai^'^^ who has conducted 
experiments using iron where the Eh values were close to the 
water-stability boundary where water is reduced to H2.^^ To 
achieve the desired low Eh values in those experiments, Rai 
found it necessary to use a specially designed vessel that 
was placed in a glovebox purged by an inert gas. Use of an 
inert-gas glovebox, had it been available, would have 
minimized oxygen in-leakage in the present experiments 
where oxygen may have diffused through the rubber stoppers 
and plastic tubing. '' Another possible oxygen source may 
have been the sample line, since it was not greased with 
silicone. 

Although the measured Eh never reached the low values 
indicated by the codes, the measured Eh and calculated Eh 
show agreement at a reaction progress of 0.0009 (Figure 24 
Curves 2 and 4 , respectively) . The measured pH of the 
system matches that of the codes at a reaction progress of 
0.0023 (Figure 24 Curves 1 and 3, respectively). Although 
one would expect the match between the experiment and 



3pr. Dhanpat Rai is a Scientist and a Technical Group Leader 
in the Biochemistry Resources department at PNNL. 



101 
calculation to occur at the same reaction progress for both 
pH and Eh, which did not occur because of the oxygen in- 
leakage, it is worth noting that both these values (0.0009 
and 0.002 3) correspond to the region where FegOs will 
precipitate. This is shown by the vertical lines in 
Figure 22 and Figure 23. This is also in the region where 
TC3O4 begins to precipitate (reaction progress of 0.0009). 
The Eh/pH in the experiment are in the stability range for 
the formation of Fe203 on a Porbaix diagram /° thereby 
indicating that the iron was oxidized and hence, the 
technetium was reduced in the experiments. 

The EQ3/6 calculations indicate that as the iron is 
oxidized to FejOs and Fe304, the technetium will be reduced 
to TC3O4 or Tc(0H)2. The intermediate oxidation state, TCO2 
(sometimes referred to as Tc02-2H20), is never indicated to 
form in these systems. 

The EQ3/6 codes contain data for Tc02-2H20 rather than 
TCO2. Cartledge^^ originally published data on TcOj, but 
has since referred to it as Tc02-2H20 and Tc (0H)4."''^^''''^ 
More recently, Meyer, Arnold, Case, and O'Kelley have 
measured the thermodynamic data for Tc02-2H20 and have found 
that it agrees well with Cartledge's original research if 
the compound is assumed to be TCO2 ■ 2H20.'^^''^^ 

Upon examining the vessel at the end of the experiment, 
the iron on the bottom of the vessel was brownish on top 
with hints of a red hue. It also had black spots showing 



102 
through the brown layer. Under the brownish top layer, the 
iron was a rich black color. The black and red colors 
observed were most likely the characteristic colors of an 
iron(III) oxide (such as hematite, Fe203) .'*^ The brown top 
layer was most likely an iron (III) oxide because the 
system's solution environment data indicated that the system 
Eh and pH were in the region where hematite (FejOs) is the 
stable phase. Even though brown is also the color of 
technetium (IV) oxides (such as Tc02-xH20) ,^^ the formation 
of this compound was not supported by the codes. 

Technetium concentration and pH results from this 
Sealed Vessel Experiment at 70 °C were consistent with the 
EQ3/6 codes, but the measured Eh was well above the 
calculated value because of oxygen in-leakage. 

In work that may be relevant to the current 
experiments, the behavior of iron in the presence of silica 
has been noted by Jantzen'^ in her studies of the leach 
rates of silica, iron, boron, and sodium from simulated 
waste glass under both anoxic and oxic conditions. In 
Jantzen's studies, the simulated waste glass was placed on 
top of the iron bars inside the vessel.^' 

Jantzen^^ showed that under oxic conditions, iron bars 
were oxidized, and aqueous iron was produced. As the iron 
oxidized, the system Eh was driven down into the stability 
range for iron silicate (FeSiOs) . The aqueous iron combined 
with the aqueous silica. As the iron was oxidized further. 



f "v « * 103 

and the silica in solution was scavenged by the aqueous 
iron, the dissolution of the simulated SRL 165 waste glass 
was enhanced to produce more silica in solution, and hence, 
more iron silicate. The Eh range for the formation of iron 
silicate under oxic conditions was in the stability range 
for TCO4", which would indicate that technetium would not be 
reduced to lower valence states and would not precipitate 
from the solution. 

For the anoxic studies conducted by Jantzen,^^ the 
leachant was deionized water and it was sparged with argon 
to deoxygenate it. The experiments were conducted in a 
glove-box continually purged with argon. The lack of oxygen 
in the system limited the amount of iron that went into 
solution from the oxidation of the iron bars. However, as 
in the oxic experiments, there was enhanced silica leach 
rates from the waste glass. The mobile silica in solution 
then reacted at the solid-solution interface of the iron 
bars and formed a protective layer of iron silicate on the 
iron surface. This used up the limited amount of dissolved 
oxygen present in the solution. Because there was no influx 
of oxygen, the reaction ended once the iron surfaces were 
covered with iron silicate. 

As in Jantzen oxic studies," the possibility that iron 
silicate may have formed in the Sealed Vessel experiments 
was considered because the vessel was glass, and the J-13 
solution does have silica in it. This is important because 



104 
the stability region on a pH/Eh diagram for iron silicate is 
in the same region where technetium would exist as TCO4' . 
Because the TcO^" in the Sealed Vessel Experiment was 
clearly reduced to lower oxidation states and the technetium 
concentration dropped BDL, the system pH/Eh could not have 
been in the stability region for iron silicate. 

The possibility of protective layers forming in the 
current experiments, as in the anoxic studies by Jantzen,^^ 
was also considered. Apparently, the rate of iron oxidation 
by the oxygen in-leakage in the Sealed Vessel experiments 
was faster than the formation of a protective layer of iron 
silicate. If it was not faster, then these protective 
layers would have prevented the iron metal from reducing the 
technetium, which in turn would have prevented the 
technetium concentrations from going BDL. 

Guppy and Atkinson^^ have conducted anaerobic 
experiments in an argon-filled glove-box where magnetite was 
put into a borosilicate glass bottle with a technetium- 
spiked solution. In these experiments, there was only a 30% 
technetium reduction, compared to the 99.8% expected had 
TCO2 been the solubility controlling compound. Guppy and 
Atkinson concluded that, under anaerobic conditions, 
technetium compounds can precipitate onto, or otherwise 
coat, the iron with a protective layer and prevent the iron 
from being further oxidized. This would prevent the iron 
from buffering the solution. 



fy '.' 



105 
Although the Sealed Vessel experiments were not like 
the anaerobic experiments conducted by Guppy and Atkinson 
because they were not conducted in an argon-filled glove- 
box, there was restricted oxygen flow. This restricted 
oxygen flow may actually be more like that expected at a 
geologic repository. Nonetheless, the possibility of the 
technetium forming a protective layer needed to be 
considered. 

For the Sealed Vessel experiments, the iron surfaces 
were not occluded because the technetium concentration 
decreased and remained BDL the entire length of the 
experiment (Figure 23). The system Eh also decreased and 
remained low the entire length of the experiment. Had a 
protective layer formed on the iron and prevented the iron 
from buffering the solution, the Eh would not have decreased 
much because of the influx of oxygen. Also, had the iron 
not been capable of keeping the system reduced, the influx 
of oxygen would have re-oxidized the technetium, which would 
have brought its concentration above detectable limits. 
This did not occur as can be seen in Figure 2 3 because the 
technetium concentrations remained near to or below the 
detection limit after the first day of the experiment. 

4.3.2.2 Sealed Vessel Experiment at 24 "C 

The Sealed Vessel System at 24 °C was previously modeled 
in Section 2.3. Figure 9 showed the technetium 



106 
concentration, and Figure 10 showed the solution environment 
versus reaction progress. For comparative purposes, the 
data in these figures have been combined with the data from 
the Sealed Vessel experiment at 24 °C. Figure 25 shows the 
technetium concentration. Figure 2 6 shows the system pH, 
Figure 27 shows the system Eh, and Figure 28 is the 
combination of Figure 26 and Figure 27. In all of these 
figures, the experimental data are plotted versus time 
(lower axis of abscissa) , and the calculated data are 
plotted versus reaction progress (upper axis of abscissa) . 

The measured technetium concentrations from this 
experiment are represented by Figure 25 Curve 1. The 
technetium concentration decreased to 2.5 x lo"*^ ppb on day 
13. By day 39, the technetium concentration was 
1.6 X 10'^ ppb, just above detection limits. The remainder 
of the samples had technetium concentrations BDL. This 
portion of the experiment is consistent with the codes 
because the codes indicated that the technetium 
concentration should drop BDL after a reaction progress of 
about 0.0009, Figure 2 5 Curve 2. 

The technetium concentration in this experiment took 
much longer to drop BDL than in the higher temperature 



107 



in 



01 

Q 



8 



o 

p IT) 

to '2 



CO >» 

•o 

< 



o 






<v 
tf) 




E 

0) 



Q 

I 
(1) 
m 






8 



18 



S 



^ 




n 



0. 
1 

0. 

C 

o 
S 

8 

c 
o 
O 

E 

3 



O 

H 

c 
o 



TO 




3 




E 




(0 




CO 








CO 




O 




HI 




0) 




^ 








? 




a 




O 









■* 




rg 








CO 








c 




<u 


OT 


E 


T3 

C 




-1 


K 


o 

Q. 


iri 


E 




o 


0) 


C) 


(/I 




(0 


b 

3 


■a 

0) 


0) 

c 



to o 
CO I- 



C\J 



(s8|ouj) paiBiidioejd pnpojd '(qdd) uojiBJiuaouoo lunjiaugoai 



a? 

3 



108 



8 
o 



o 




m 
in 



!* 



9 



X 

o 



CO 
(V 

£ 
o 



<u 

£ 

o 



o 

O 



RI 


^ ^ 


-) 


CM 


f- 


0) 


w 


3 


^ 


o 


O 




Ui 





o 

CO 



CM 



o 

CM 



>. 

CO 

E 



O M- (O CM •^ 



OOOOt^UJUlTfOOtNi- 



n 



Hd 



u. 



8 

d 



109 



o 



o 


s 


rn 


■o 


E 














CO 


m 


frt 


ro 


>. 




33, 






T3 






* I 



#1 



* 



8 



i? 



? 



8 

o 



v> 

p 



c 
o 

t !2 



^ 


13 

























Q. 





0) 



o 
o 



8 

d 



E 

>. 
w 



ra 



(g 
n 
O 

LU 



o 



* 



II 
o; 






8 






W 
(0 

0) 

E 



o 




8 8 
00 h- 



o 



2 3° 
000 
CO (SI i^ 



8 8 



O 
O 
01 



O 



(ALU) 43 



'E 
3 

> 

ro 

CO 
c3 

3 

o 



•o 

0) 

w 
'0. 

■3 

0) 



c 
g 

ro 

3 
E 
w 

(O 
CO 

O 

UJ 



T3 

CO 



CM 

ro 

c 
e 

E 

-c 

& 

X 

lU 

w 

o 

O 
Q. 



> 

00 
W 



CM 

£ 

3 
Oft 

iT 






110 



in 

CO 

o 
o 



CO 

8 

d 



(Nl 

8 

d 



/-v 


S 


N 


o 








< ) 


V) 




V) 




a> 








O) 




o 








Q. 




c 




o 








C) 


u> 


CD 


R 


tr 


o 



i 

+ 



•4 

•4 



o o 






i 

CO 



•*? 



in 
o 
o 
o 



E 



3 

E 


<o" 


<n 


Q) 


(O 


E^ 






CO 


f 1 


O 





1) ^^ 

(/) ;r 

<U (U 

> e 

0) O 

(0 "^ 

<u 
« 

O 



X 
a. 




I 



(5 ^ ^ 



c 


c 


E 




re 

0) 








<o 




1 


^ 


O 


B 


I 


LU 


UJ 


m 


m 


U 


£ 


T 


^ 


T 




UJ 


Q. 


lij 


Q. 



* f 



o 










0) 








CO 




w 








o 




3 


^ 


■D 
S 




Q. 




■o 




01 








CO 






o 


Q. 


in 


O 




0) 








CL 



I 
o 



CO a; 



lU 



§ 



o 

U 






LU 

T3 

c 

CO 

I 

Q. 

c 
o 







3 






O 






CO 






c 






o 


in 






CO 




3 

F 










^ 


w 




CD 


CO 


8 


2, 


CO 


(U 


U 




F 


111 




H 


0) 
T3 


in 




c 


CM 




CD 

rg 

TO 


in 




E 

•c 

X 

LU 



(A) 43 puB Hd 



!1> 



> * ■ r 






111 



experiment. it probably took longer than 13 days^ for the 
technetium concentration to drop to BDL because of slower 
reaction kinetics. Another possible contributing factor was 
that the technetium took longer to diffuse through the 
solution to the iron surface on the bottom of the vessel 
where a reaction could occur. Both the diffusion of 
technetium and reaction kinetics are slower at lower 
temperatures. 

For the first 39 days, the pH of the system also 
corresponded well with that predicted by the EQ3/6 codes, 
Figure 2 6 Curves 1 and 2 , respectively. However, the pH of 
the system dropped when sodium chloride (NaCl) was added, 
and the temperature was increased to 70=0 (the salt addition 
and temperature increase will be discussed momentarily) . 
The simulation did not model this temperature increase or 
salt addition, and therefore, no comparisons between the 
experimental and calculated pH were made after day 39. 

The 0.88 unit drop in pH at day 39 was expected as a 
direct effect of the 46oc temperature change. This was 
verified by using the Nernst equation. Equation (5) with 
either iron metal or Fe^ in equilibrium with hematite 
(Fe203) . Hematite was used because the pH and Eh values 



^^"t^^9%%s'lls%f^^^^^^^^ next data point 

conservative with reaa?d to ^hf f H^iits. To be 

technetium concentrJ?Ion to drSp^B^L "t^.Tt '?" "^^^ 

former point was used. ^ ' ^^^ '^^^^ ^°^ ^^e 



112 
measured in this experiment are within the region where 
hematite is stable. The sample collected after the salt 
addition showed the technetium concentration BDL, as would 
be expected with the pH/Eh condition of the solution at that 
point. This is illustrated by the calculations by following 
Curve 3 from the experimental data over to the calculated 
data, and then following Curve 4 down to the bottom of the 
graphs. Both Figure 2 6 and Figure 27 indicate that the 
technetium precipitated as TC3O4. 

The measured Eh of the system was much greater than 
expected due to oxygen in-leakage. On day 39, sodium 
chloride was added to the solution in an attempt to lower 
the Eh by increasing the corrosion rate of iron. At the 
same time the temperature was raised to 70°C. This 
combination caused the Eh of the solution to drop 
substantially, and it eventually leveled out near -150 mV. 
This was a lower value than in the Sealed Vessel Experiment 
at 70°C without NaCl, Figure 23 Curve 1, but it was above 
the values calculated by the codes, Figure 27 Curve 2. 
Eleven days later while attempting to lower the Eh more with 
a second salt addition, additional air was inadvertently 
admitted into the system. This was quite noticeable by the 
jump in the Eh at day 50. 

The measured Eh probably never reached that predicted 
by the codes because oxygen leaked into the system for the 
same reasons as stated in the Sealed Vessel Experiment at 



113 

70°C (refer to Section 4.3.2.1). The Eh in this experiment 
was also substantially higher than in the Sealed Vessel 
Experiment at 70<>C, Figure 23 Curve 1. This was because the 
oxygen leaking into the system reacted more slowly with the 
iron at the lower temperature. 

The iron in the bottom of the reaction vessel contained 
the same brownish top layer with hints of red, and it was 
black underneath the top layer. This was most likely an 
iron (III) oxide as described for the Sealed Vessel 
Experiment at 70°C (Section 4.3.2.1). 

Just as with the Sealed Vessel Experiment at 70°C, data 
from this experiment, other than the solution Eh, were 
consistent with the EQ3/6 codes. 

4.3.3 FT-4C Iron Powder Experiments 

The FT-4C system was used to conduct three ''tc Iron 
Powder experiments. The High-flow and Low-flow experiments 
were conducted at 24 °C, and the third experiment was the 
iron Powder at 85oc. m Section 4.2, it was determined that 
this system was retaining technetium in the system's 
plumbing and releasing it at a later time. Therefore, no 
direct conclusions can be drawn from these experiments. 
They are included for completeness. . 



<« 






■ ■• .'f 






'W..,./^ „,,,._-...- ^^4 

4.3.3.1 High-flow Iron Powder Experiment 

The High-flow Iron Powder Experiment used the FT-4C to 
pump a 1500 ppb 9'Tc-spiked solution through the iron-filled 
reaction column at 50 mL/day, and 100 mL/day. Then the flow 
was stopped such that the solution was in static contact 
with the iron. Data for this High-flow Iron Powder 
Experiment are shown in Figure 29. After the initial sharp 
decrease from the starting ''tc concentration of 1533 ppb, 
the concentration gradually decreased further to a minimum 
of 0.15 ppb. At the end of the third day, the flowrate and 
pressure were turned off to tighten a loose bolt on the 
reaction column. Although it only took 10 minutes before 
the flow was re-started and the system pressurized, this 
interruption apparently caused the technetium concentration 
to increase for 1 h before starting to decrease again. 

Two and a half hours after tightening the bolt, the 
flowrate was increased to 100 mL/day for 5 h to determine if 
the shorter contact time between the iron and the solution 
would allow the technetium concentration in the effluent to 
increase again. The technetium concentration did not 
increase as was expected. Instead, 1 h after the flowrate 
change, it had actually decreased to about what it was 
before tightening the bolt. Thus, the change in flowrate 
appears to have had little effect on the technetium 
concentration. 



115 




(A 

CO 

E 



(qdd) uouBJjuaouoQ LuriDauijoai 



c 

Qi 

E 

'u- 
0) 
Q. 
X 

LU 

u- 
CD 

1 

o 
CL 



o 



if 
o 



CM 



116 
Next, the flow was stopped to determine if the 
technetium concentration would decrease further because of 
longer contact time with the iron. Although this portion of 
the experiment was labeled no-flow, there was a very small, 
insignificant flow as the samples were collected. During 
this period of no flow, the technetium concentration 
increased slightly instead of decreasing. In hindsight, 
this increase appears to be from technetium that was 
deposited in the system's plumbing during an earlier 
experiment as described in Section 4.2. 

It should be noted that some effluent was observed on 
the outside of the column when the FT-4C shielded doors were 
opened to tighten the loose bolt on the reaction column. 
The liquid that leaked was milky white, and it left a white 
solid in the reaction column holder. This solid may have 
contained a precipitate of technetium, or it could have been 
due to the J-13 evaporating. Although it is possible that 
this solid was caused by the liquid evaporating, it appears 
more likely that technetium was precipitating because (1) 
the reaction column was not heated (which would have 
increased the rate of evaporation), (2) the solution was 
milky white, and (3) similar white precipitates were 
observed in both of the stannous chloride experiments where 
evaporation did not occur, as will be discussed in 
Section 4.5. 



'.An ' ;i.i -. 



■^ V 



r ( 1 



117 
This precipitate was probably TC3O4 because other 
compounds, such as TcOj (sometimes referred to as Tc02-2H20 
or Tc(0H)4), have been reported to be a brownish 
precipitate'^^ and possibly even a black sludge and brown 
deposit /2,43,46 j^Q trace of a black or brown residue was 
observed. Also, Tc02-2H20 has a solubility of 4 . 4 ppb,^i 
which is much greater than was observed in the experiment. 
The codes indicate that TC3O4 will form, not Tc02-2H20. No 
data for the color of the precipitates for TC3O4 or Tc(0H)2 
exist. However, the solubility limit for Tc(0H)2 is BDL and 
because technetium was always detected in this experiment, 
the precipitate would not have been Tc(0H)2. As will be 
discussed later, a similar white precipitate was also 
observed in the stannous chloride experiments. 

4.3.3.2 Low-flow Iron Powder Experiment 

A second iron-powder experiment was conducted. This 
time the flowrate was lowered after 3 days to determine if 
allowing longer contact time between the iron and technetium 
would decrease the technetium concentration to lower values 
than in the High-flow experiment. During those first few 
days, the two experiments were conducted at the same 
flowrate, and the results were similar in that there was a 
large decrease in technetium concentration (Figure 30) . 
However, the technetium concentration did not change much 
during the 18-day period at 10 mL/day. Perhaps the minimum 



I 

^— C . i- 118 

technetium concentration of about 0.2 ppb in both the High- 
flow and Low-flow experiments was due, in part, to the 
release of technetium previously retained by the system, as 
described in Section 4.2. 

4.3.3.3 Iron Powder Experiment at 85 "C 

The next iron powder experiment was an attempt to 
increase the reaction kinetics by increasing the temperature 
to 85°C. The flowrate was held constant at 50 mL/day. The 
results, shown in Figure 31, were similar to those of the 
two lower temperature iron experiments. That is, the 
technetium concentration dropped to a minimum of a few 
tenths of a ppb. Then it slowly increased and by the end of 
the experiment was 5 ppb. 

The minimum technetium concentrations in all three 
FT-4C experiments were in the same range as the minimum 
concentrations found in the Blank-run Experiment described 
in Section 4.2. This suggested that the minimum 
concentrations in the FT-4C experiments were an artifact of 
technetium being retained in the plumbing downstream from 
the reaction column and then re-released. This was not a 
true indication of the effectiveness of the iron. This was 
a clear indication that a different system was needed to 
investigate iron as a reactant to reduce the technetium 
concentrations. Therefore, all flow-through experiments 
conducted after these three used the Bench-top system. 



fiyf'f: f 



».' 






119 




& 



(qdd) uouBJ^uaouoo LunijauLjoai 



O 

I 

I 



120 




CO 






(qdd) uojiBJjuaouoo oinjiaugoai 



O 



& 

3 

u. 



121 
4.3.4 Iron Powder Summary and Conclusions 

All the experiments conducted using iron as a reactant 
had a substantial drop in the technetium concentration. The 
technetium concentration in the Bench-top ''tc experiment 
dropped from a starting concentration of 35.8 ppb to BDL 
(0.4 ppb). The Bench-top ^^"^c experimental results showed 
that the technetium concentration dropped from a starting 
concentration of 9.4 x lo''^ ppb to approximately 
1 X 10"^ ppb. The two Sealed Vessel experimental results 
also showed that the iron reduced the '^""Tc concentration 
from 2.3 X 10"^ ppb to BDL, i.e., less than 1 x lo"'' ppb. 

Before comparing data from the Bench-top experiment to 
that from the Sealed Vessel experiment, recall that in the 
Sealed Vessel System (1) the amount of technetium was 
limited rather than being replenished as in the Bench-top 
system where the spiked solution was pumped through the 
iron, (2) the amount of oxygen available to the system was 
substantially less than in the Bench-top system, (3) the 
surface area was substantially less because the solution was 
not pumped through the reaction bed as in the Bench-top 
experiment, thereby leaving the top layer of the iron as the 
primary surface area available for the technetium, and (4) 
the time the technetium was in contact with the iron was 
days rather than the 1.5 h it took to pump the solution 
through the reaction column in the Bench-top system. 



'^At'-:"''fh*' : .' , i' ^^-PO 



122 
Now, comparing the results of the Bench-top '^"^c 
experiment with the Sealed Vessel Experiment at 24 °C, we 
know (1) it took longer than 13 days in the Sealed Vessel 
experiment for the iron to decrease the technetium 
concentration BDL, and (2) the technetium solution was in 
contact with the iron for only 1.5 h in the Bench-top 
system, but the technetium concentration was 7% lower 
(1.8 X 10''') than that in the Sealed Vessel experiment 
sample collected on day 13 (2.5 x lo""^ ppb) . Therefore, it 
is evident that in the Bench-top Experiment, some phenomenon 
other than just reduction occurred to lower the technetium 
concentration in the sample effluent so quickly. 

As described in Section 4.3.1, the technetium was 
likely sorbed by the iron, causing the rapid drop in 
technetium concentration. This can occur more quickly when 
the solution is pumped through the bed of iron powder than 
when the iron powder lays at the bottom of a large volume of 
solution under static conditions. Since the technetium was 
unlimited in the Bench-top experiment, it did not drop BDL 
as in the Sealed Vessel experiments, and it eventually 
started to increase from its minimum concentration of 
(1 X 10'^) . 

The 95m-pc concentrations in the Bench-top Experiment 
were markedly reduced, but not to the values as low as 
calculated by the EQ3/6 simulation. This reduction in 
technetium concentration was due in part to technetium 



"1.1 






123 



i \lt ' - » ' 

sorbing on the iron surface (which the codes cannot model) , 
as well as the technetium being reduced by the iron. Once 
the iron sorbed the technetium, its surface became partially 
passivated as was indicated by the slight rise in technetium 
concentration in the latter half of the experiment as well 
as by the arguments discussed in Section 4.3.1. 

Comparing the Sealed Vessel Experiments at 70 °C and 
24 °C shows that the increase in temperature substantially 
reduced the time it took to decrease the technetium 
concentration BDL. In the 70 °C experiment, the technetium 
concentration dropped BDL within the first day. However, it 
took greater than 13 days for this to occur in the lower 
temperature experiment, most likely because the reaction 
kinetics of iron are slower at room temperature. A chemical 
reaction (technetium reduction) , rather than technetium 
sorption, is supported by the visual examination of the iron 
after the experiment ended. The iron on the bottom of the 
vessel was brownish on top with hints of a red hue. It also 
had black spots showing through the brown layer. Under the 
brownish top layer, the iron was a rich black color. The 
black and red colors observed are most likely the 
characteristic colors of an iron (III) oxide (such as 
hematite, FejOj).^^ Technetium sorption, as a mechanism of 
the slower drop in technetium concentration at 24 °C, also 
seems to be ruled out because the associated passivation 
layer would be expected to form slower at the lower 



124 
temperature. Slower formation of a passivation layer would 
have allowed the technetium concentration to decrease 
faster, not slower. Because the technetium concentration 
was measured to decrease more slowly at the lower 
temperature, technetium sorption must not have been 
involved. The iron can sorb the technetium until the 
sorption sites are occluded by being filled. A separate 
mechanism but with the same effect is the development of a 
passivation layer. Both mechanisms, occlusion of sorption 
sites and development of a passivation layer, would limit 
the amount of technetium concentration reduction and can 
both occur at the same time. 

Diffusion may have been another factor that contributed 
to the slower decrease in the technetium concentration in 
the lower temperature experiment. Technetium diffusion 
through the static solution to reach the iron surface on the 
bottom of the vessel where the reactions occurred would be 
expected to be slower at the lower temperature. 

The technetium concentration decreased in all three of 
the FT-4C experiments, but not to the levels calculated by 
codes. Because the minimum technetium concentrations were 
within the same range as the Blank-run Experiment, it is 
likely that the remaining technetium was the result of 
technetium being retained and then re-released by the 
system. Therefore, no conclusions can be made regarding 



♦ J I 






125 
iron's redox effect on technetium for the experiments 
conducted using the FT-4C. ^" 

The technetium concentrations remained at low levels 
for the duration of all seven of the experiments (two Bench- 
top, two Sealed Vessel, and three FT-4C experiments) . The 
experiments continued up to 56 days as in the Sealed Vessel 
Experiment at 24 °C. Maintaining low technetium 
concentrations for geologic time periods in a repository 
environment would require the continued presence of iron in 
an active (non-passivated) form. Further work is needed to 
determine whether iron can be expected to remain active for 
■ extended (geologic) time periods. 

4.4 Copper Experiments 

The four copper experiments listed in Table 4 were 
conducted. The Bench-top Flow-through System, Figure lib, 
was used to pump '^mTc-spiked J-13 through a reaction column 
containing copper powder. The details of the Bench-top 
Copper Powder Experiment are presented in 4.4.1. The other 
three copper experiments used the FT-4C system. Figure 11a, 
to pump Tc-spiked solutions through reaction columns 
containing copper shot. 

There are some uncertainties associated with the data 
collected from the FT-4C because the system was retaining 
technetium in the plumbing downstream from the reaction 
column and re-releasing it during these experiments 



t . -, ''■ ■ ■ -" ■ 

" ^,. . , . -. ..l\- I h- ^ 126 

(Section 4.2). However, the data from these experiments are 

credible because the technetium concentration in the samples 

were orders of magnitude greater than that detected in the 

Blank-run Experiment. The details of the three FT-4C copper 

experiments, are presented in Sections 4.4.2.1 through 

4.4.2.3. Despite the uncertainties with the FT-4C, there is 

no question that the technetium concentrations remained 

essentially unchanged when the solution was passed through 

the reaction column. 

Because the FT-4C experiments were conducted using 

copper shot, copper powder was used in the Bench-top 

experiment to determine whether the larger surface area 

would reduce the technetium concentration. The Bench-top 

experiment is discussed in Section 4.4.1, and the FT-4C 

experiments are discussed in Section 4.4.2. 

4.4.1 Bench-Top Copper Powder Experiment 

The Bench-top Copper Powder Experiment was conducted 
after the three FT-4C experiments. It is presented first 
because of the problems with the data from the FT-4C as 
described above. However, the trends in the data from all 
four experiments (three FT-4C and one Bench-top) were 
consistent, i.e., the technetium concentrations were 
essentially unchanged except for an initial drop or rise. 

The Bench-top Copper Powder Experiment had a ^^'^c- 
spiked solution pumped through a reaction column loaded with 



127 
copper powder. The other details of the experiment are 
shown in Table 4 . This Bench-top experiment used copper 
powder rather than copper shot because the effect of surface 
area on technetium retention was being investigated. In 
addition, this experiment was designed to determine if the 
initial drop/rise in technetium concentration was 
reproducible and to determine why it occurred in the 
experiments, but not in the computer simulations. 

The data for this experiment are shown in Figure 32 
Curve 1. Curves 2 and 3 in Figure 32 are the results of the 
FT-4C Copper Shot experiments at 24 °C and 85 °C, respectively 
(see Section 4.4.2), and are included for comparative 
purposes. This experiment was conducted before redoing the 
computer simulations as mentioned in Section 2.3. Those 
early computer simulations used both copper and technetium 
metals as reactants. The technetium metal reactant allowed 
the simulation to continue even when the system was 
saturated with copper.* Because technetium can exist up to 
the (VII) valance state, the simulations that used 
technetium metal as a reactant calculated much lower Eh 
values than the simulations that only used copper metal as a 
reactant. The lower Eh values in the simulations containing 
technetium metal reactant indicated the solution was highly 
reducing, the technetium precipitated out of solution, and 



recall that when a system is saturated with a reactant, it 
IS in thermodynamic equilibrium, and any additional reactant 
added will have no effect on the equilibrium of the system. 



■ •»' 






128 




CO 



c 
D 



c 

E 
'(_ 



Q. 
X 
UJ 

o 

x: 
W 

i_ 

(U 
Q. 

Q. 
O 

o 

o 



O 



(qdd) uojjBJiuaouoo ujniiaugoai 



129 
the technetium concentration dropped below 1 x lo'^ 
(technetium concentrations are not indicated on those 
earlier graphs but they were indicated to drop below 
1 X 10"^) . Based on those simulations, the starting 
concentration used in this experiment, 9.4 x lO"^ PPb, would 
have resulted in a 4 order of magnitude drop in technetium 
concentration to reach BDL (1 x lo'' ppb) . More recent 
computer simulations, which used copper metal as a reactant 
and technetium concentrations similar to those in the 
experiments, indicated that the technetium concentration 
would be reduced to 3.4 x lo'^ ppb. In these latter 
simulations, the simulation ended once the system was 
saturated with copper metal but before the technetium 
concentration dropped below 1 x lo''' ppb. The technetium in 
these latter simulations was entered as the pertechnetate 
ion. Because the technetium was in its highest oxidation 
state (VII) , it was not able to drive the Eh of the system 
lower, and therefore once the copper was in equilibrium with 
the system, the simulation ended. Had the latter 
simulations been conducted prior to the experiments, a 
higher starting technetium concentration would have been 
used. 

Figure 3 2 Curve 1 shows that the technetium 
concentration initially dropped from the 9.4 x lO"^ ppb 
starting value to a minimum of 1.5 x lo"*^ ppb within 3 h. 
Then it returned to essentially the starting concentration. 



'■ r';*,* > ' i*^ -^ ■ ^ 



After the initial drop in technetium concentration and 
return to the approximate starting value, the concentration 
decreased to 7.3 x lo"'^ ppb, a 21% decrease from the 
starting concentration. Although this decrease was quite 
small, it was larger than the total uncertainty of about 3% 
related to the measurement errors. The initial decrease to 
1.5 X 10'^ ppb was possibly caused by sorption. After the 
sorption sites were saturated, the technetium concentration 
returned to and remained essentially unchanged from its 
starting value. This initial decrease could also be the 
result of the technetium being reduced by the copper, and 
the rise back to the starting concentration the result of a 
passivation layer developing. Sorption and passivation 
layers are two separate mechanisms that give the same 
effect. They can occur alone or at the same time. The 
passivation layer that developed in this experiment was most 
likely cuprite, which is a protective film that forms on the 
copper and is responsible for not allowing the copper to 
oxidize further. ^° '. 

The comparison of the Bench-top experimental results 
with the computer simulation of this experiment is shown in 
Figure 33. The experimental results are plotted versus time 
(lower axis), and the computer simulation results are 
plotted versus reaction progress (upper axis) . 

The lowest calculated technetium concentration was 
3.4 X 10-^ ppb (Figure 33 Curve 2), which is much higher 



131 




(A 

C 

o 



o 

o 
o 



Q. 
O 

Q. 



5 "C 

II 

O) CO 

*- o 

"O T3 

-1 

3 CO 

i2 w 

TO c 

Qi g 

^ 00 

O) i; 

i^ g 

. o 

c c 

o o 

.s o 

CD i_ 

3 03 

E Q. 

CO 8 

JO -o 

CO e 

o ■" 

UJ E 

03 .E 



(saiouj) spnpojd pajBiidpaJd puB (qdd) jaddoo puB uinnauijoai 



^ 0) 




and 
echn 








li 

.E TO 




1- — 




0) 3 




Q. O 




X TO 




UJ O 








1 •" 




o «■ 




Q. ^ 




^ E 




0) -^ 


N 


Si 

o 2 


if? 


O - 


Z. cc 


CD 


Q. -^ 


> 


O = 




li 


■o 


B 




" ir 


o 


<u 9- 


Q. 


m 2 


£ 


S w 


CO 


£ =) 


« 


<- m 


CO 


O 0) 
c > 


1^ 

0) 

c 


.<" ^ 


E 


1_ ■*-' 




mpai 
plot 


■o 


o 2 


2 


O CO 


Q. 


CO 




CO 




£ 




3 




a> 




il 





f\f\.r 



132 
then the experimental minimum of 1.5 x lO'* ppb (see note on 
the figure) . The codes followed the individual packet of 
solution (see Section 2.3.2). As the reaction progressed 
and the solution environment became less oxidizing because 
of the oxygen being depleted, the copper metal reactant was 
first oxidized to Cu(II) as tenorite (CuO, Curve 4). As the 
oxygen was depleted more, the copper reactant was only 
oxidized to Cu(I) as cuprite (CujO, Curve 5) and delafossite 
(CuFe02, Curve 6) . The solution became even less oxidizing 
as it progressed further through the copper in the 
simulation. The technetium in the solution was reduced to 
TC3O4 (Curve 7) at a reaction progress of 0.00065. At this 
same reaction progress, essentially all the dissolved oxygen 
was depleted. 

The simulation was terminated after all the dissolved 
oxygen in the packet of solution was exhausted and copper 
was indicated to be the most stable compound (curve 8) . 
Even though Tc04', and hence oxygen, was available in the 
simulation, it did not oxidize the copper reactant. The 
technetium concentration did not decrease below 
3.4 X lo'"* ppb because the simulation was terminated once 
copper metal was indicated to be the most stable copper 
phase, but before all the technetium precipitated. 
Experimentally, the technetium concentration did not change 
much after the initial drop. Based on a recent computer 
simulation, little change would have been expected because 









4 " . ' 



133 
the lowest calculated technetium concentration was only 
about 3 times lower then the experimental starting 
concentration. Therefore, the approximate agreement between 
the experiment and the simulation (Figure 33) may have been 
a coincidence of the low experimental starting 
concentration. 

To determine if the simulation containing copper would 
decrease the technetium concentration BDL or at least as low 
as the minimum obtained in the experiment, agueous Cu* was 
entered as the reactant instead of copper (metal) . Copper 
metal was suppressed from forming. The results of this 
computer simulation are shown in Figure 34. The initial 
drop in the measured technetium concentration (Curve 1) 
corresponds with the same calculated technetium 
concentration (Curve 2) at a reaction progress of 0.0017 
(follow Curve 6 from the experimental minimum to the 
calculated value) . At that reaction progress, cuprite, 
delafossite, and TC3O4 began to form. 

According to the agueous copper simulation, the 
technetium concentration should have continued to drop as 
the copper was oxidized to cuprite and tenor ite. This did 
not occur in the copper metal simulation because there was 
no more oxygen available to continue to oxidize the copper, 
and copper metal was indicated to precipitate. TC3O4 was 
never allowed to completely precipitate in the copper metal 
simulation because there was no further reaction once all 



■fe-'. 



134 



eo 

o 
o 



o 
o 



8 

o 
o 



if) 
o 
o 




0) 

■~ .^ o X 

a n c Q. S 9- 

ro o 4) 3 <i> o 

O H t" O Q I- 



r 



s I 



t 

w 

w 

o 



o 
o 



CO 
o 
o 

d 



CM 

o 
o 



E 
5-2 



g S 



"8 



J= c 

(D O 

3 O 
o 

O 



O 



O 

o 



o 
o 
o 



" o 
o 



CO 

+ 
UJ 



+ 
UJ 



LU 




oo 



(O 



m 



(0 

E 



- CO 



CN 



(S9|0U]) 

spnpojd paiBijdioajd pue (qdd) uojieJiuaouoQ luni^auMoai 



1i 

3 
O 

s 

1 

3 
CO 
(0 

o 



o 




O 




CO 




3 




O 




O 




3 




O" 




< 




a> 




c 


TJ 


(/I 


0) 


D 


j- 






c 


n 


o 






(/) 


TO 


T3 


3 


3 


fc 


o 


w 


n 


(O 


■D 


C) 


0) 


O 


CO 


III 






n 






r 


o 




0) 






■o 


n 


c 




m 


u 




11 






t- 


m 


(1) 




E 


3 
O 


•c 


o 


X 


u 


III 


c 


i- 


(0 


a> 


w 


1 


c 
o 


o 


♦-I 


a. 


iu 


^ 


p 


o. 


o 
O 


8 


Q. 
|5 


E 

3 


f- 


0) 


U 


c 




■Jj 


m 


5J 


■t 




CO 




0) 








3 




o> 





the oxygen was gone, and the system was thermodynamically 
driven to precipitate copper metal. The codes did not 
indicate that there will be a point where the technetium 
will preferentially precipitate over the copper metal. 

The simulation using copper as an aqueous solution did 
not model the experiment; it was just used to determine the 
solubility-controlling compound for copper in solution. The 
technetium concentration did not begin to decrease until 
cuprite and delafossite (Figure 34 Curves 4 and 5) have 
began to form and precipitate. The codes do not model 
passivation layers, but cuprite was most likely the 
solubility-controlling copper compound in the experiment. 
Cuprite could form under the current solution environment 
conditions, it is a protective film,'^° and the reduction of 
the technetium concentration in the experiment was 
temporary. Once this protective layer formed in the 
experiment, no further reduction occurred. 

The effect of surface area was also studied by 
comparing the experiments that used copper shot and copper 
powder. However, because two variables were changed, the 
surface area and the initial technetium concentration, 
comparison is difficult. As will be shown, the FT-4C Copper 
Shot Experiment at 24 °C, Figure 3 6 Curve 1, reduced the 
technetium concentration from 1.3 x lo^ ppb to 
1.1 X 10 ppb, or 200 ppb. As seen in Figure 32 Curve 1, 
the copper powder reduced the technetium concentration from 



.:./-■- 136 

' ■ ■ ^% ' " 

1 X 10'^ ppb to 1.5 X 10"^ PPb. Therefore, the copper shot 
sorbed substantially more technetium than the copper powder. 
Whether this would be true if the same starting 
concentrations were used is uncertain. However, in both 
cases, not all the technetium was removed from solution, 
regardless of surface area available. Once the passivation 
layer developed on the copper surface, further reaction of 
the copper and solution was prevented, and the technetium 
concentration returned to near the starting values. 

4.4.2 FT-4C Copper Shot Experiments 

The FT-4C was used to conduct three copper shot 
experiments. This system was discovered to be retaining 
technetium and releasing it at some time later. In spite of 
this problem the data obtained from these experiments are 
credible because the technetium concentration in the samples 
was orders of magnitude greater than the 14 ppb of 
technetium detected in the Blank-run Experiment 
(Section 4.2). 

4.4.2.1 DIW Copper Shot Experiment 

The FT-4C system was used to conduct the distilled, 
deionized water (DIW) Copper Shot Experiment. This was the 
first experiment conducted, other than the readiness 
experiments, and DIW was inadvertently used in place of 
J-13. The technetium concentrations (axis of ordinates) 







137 



(qdd) uoijBJjuaouoo oinuauijoai 



£ 

3 

il 



138 
versus time (axis of abscissae) from the DIW Copper Shot 
experiment are plotted in Figure 35. The flowrate was 
changed from 25 mL/day to 50 mL/day, 12.5 mL/day, 50 mL/day, 
and finally mL/day on days 1.1, 2.2, 2.9, and 3.1, 
respectively. The flowrate was increased after 1.1 days 
because the initial flowrate was inadvertently set at half 
the flow of the expected water flux through the 
repository. ^'^^ The flowrate was decreased later in the 
experiment to determine if longer contact time between the 
copper and technetium would allow the copper to reduce the 
technetium. The flowrate was increased again to determine 
if decreasing the contact time between the copper and 
technetium would cause the technetium concentration to 
increase. Then the pump was finally turned off on day 3.1 
to determine if the longer contact time between the solution 
and the copper would cause the technetium concentrations to 
decrease. 

The reliability of most of the data collected during 
this experiment (Figure 3 5 Curve 1) is questionable. The 
volume of column effluent used to prepare most of the 
samples for analysis with the LSA was very small, on the 
order of 0.009 mL. The uncertainty associated with this 
volume measurement (> 3 3%) makes the data questionable. The 
data are included in this report only for completeness. 
Five of the samples from this same experiment had 
volumes greater than or equal to 0.1 mL. They are plotted 



t ■ V 



. » ."> it.'—-. ^ 

139 

on Figure 35 Curve 2. Uncertainties associated with these 

samples were considerably less then with the small samples, 

and the data show the "tc concentration essentially did not 

change during the entire 8 days this experiment was 

conducted (Figure 35 Curve 2) . 

4.4.2.2 Copper Shot Experiment at 2 4 °C 

The above experiment was repeated using J-13 instead of 
DIW. Figure 3 6 Curve 1 shows the technetium concentration 
versus time obtained from the "tc J-13 Copper Shot 
Experiment at 24 °C. These data are consistent with the 
previous experiment in that the technetium concentration 
remained essentially the same throughout the experiment. 
The initial drop in concentration within the first 2 h of 
the experiment was followed by a return to, and then above, 
the starting concentration. This may have been caused by a 
passivation layer developing on the copper surface. The 
initial drop in technetium concentration could also be the 
result of the copper sorbing the technetium until the 
sorption sites were saturated. It was not likely caused by 
the system first retaining and then re-releasing technetium 
downstream from the reaction column (refer to the Blank-run 
Experiment in Section 4.2). The maximum technetium 
concentration found in the Blank-run Experiment, 14 ppb, is 
far less than the increase of 132 ppb in the Copper Shot 
Experiment at 24°C. 



140 




O 

e 

to 

00 

CO 

O 

e 

(A 
O 

E 

•c 



o 
(0 



(qdd) UOIIBJJU90UOO LunjjauMoai 



o 
o 

CO 



O 

I 

& 



141 



I 

Q. 



>J-<a 



o 

E 
a. 




•0< 



LU 



o 



o 



O 

3 



CD 
O = 

uj 2 
2 '^ 

05 O 

c c 

E 8 
8 E 

a> 

o o 
2 ^ 



w 




(U 




o 




3 




■n 




<i) 








i- 




(U 




a 




Q. 




o 




(1 


■^ 


11 


o 


^ 


fV 


w 


1- 


TO 


o 




■4— • 


n 


E 


n 








CO 


<u 




(- 


LU 


£: 


O 


C5 




<D 



o 
o 



10 

w 

03 

o 



% 

a 






0) ±; 
o (u <« 



*-* <y 7/ 

O O 0) <u ^ OX 
KI-I-QOO Q. 



I: 



Hia 



iiiilin r Bun i-i -i.iH..^ I 



O 

as 

§ ^ 

o 

o 

o 



(O 

o 
o 
o 



o 
o 

o 

d 



CO 

o 

+ 
OJ 



o 

+ 
lU 



o 

I 

UJ 



CO 

o 

UJ 



o 

I 

UJ 



1^ 
o 

I 

LU 



o 

UJ 



eg 

o 
o 

o 



o 
o 
o 
o 



III 



UJ 



CD 

I- 



o 

3 

o 



o 

o 

E 

T3 

c 

CO 
V) 

c 
.2 

2 

c 

s 

c 

8 

E 

c 
o 

■o 

3 

o 



CO 




o 




0) 




£1 




\- 




vi 








tz 




o 




E 




■c 




fl) 


*-» 


Q 


c 


V 


CD 


UJ 


t3 




CD 


o 


a) 


j= 


^ 


W 


CD 


L_ 


(/^ 


Q) 


m 


Q. 




Q. 


0) 


O 


n 


O 


a. 


0) 


R 


£ 










P 


o 




o 


^ 






CD 


o 


3 


u. 


E 


?^ 


CO 


CD 


CO 


T3 


CO 


F 


() 


L: 


UJ 


^ 


r- 




<o 





(s8|ouj) spnpojd p!|os '(ujcld) uojjBJiuaouoo Lunuauijoai 



£ 



14 2 
Figure 37 shows the technetium concentration and solid 
product minerals indicated to precipitate by the simulation 
of this experiment. Curve 1 represents the technetium 
concentration in parts per million (ppm) . This is the only 
set of concentrations in this figure; all the other lines 
represent the number of moles of solid products 
precipitated. The calculated technetium concentration was 
decreased to 1.9 x lo"^ ppm (1.9 x 10'^ ppb) . However, this 
did not occur in the experiments. 

The first solid copper product indicated to form by the 
simulation was tenor ite (Curve 2) . Tenor ite continues to 
form until a reaction progress of 1.68 x lo"^ is attained. 
The solid copper (I) oxides are the most thermodynamically 
favored copper compounds to form at this same reaction 
progress. The copper (I) oxides indicated to precipitate are 
cuprite (Curve 3) and delafossite (Curve 4) . The first 
technetium compound also began to form at this same reaction 
progress (TC3O4 represented by Curve 5) . The technetium 
concentration, Curve 1, began to decrease as expected once 
TC3O4 (Curve 5) began to precipitate. Cuprite, delafossite, 
and TC3O4 continued to form as the reaction progressed. It 
is at a reaction progress of 1.74 x 10"^ that copper metal, 
Curve 6, was shown as a the most stable copper product. 
Nothing further happened to the copper reactant because no 
more oxygen was available to oxidize the copper, and 



j.r-:. ■;■•■■' 7-^ :\.[ H 



143 
copper (0) metal was the most thermodynamically favored 
compound. 

In general, the technetium concentrations did not vary 
much in this experiment. It was suspected that slow 
reaction kinetics might be responsible. Therefore, an 
elevated-temperature copper shot experiment was conducted to 
determine if indeed slow reaction kinetics was the cause for 
the discrepancy between the calculation and the experiment. 

4.4.2.3 Copper Shot Experiment at 85 °C 

Figure 3 6 Curve 2 shows the variation in technetium 
concentration versus time from the Copper Shot Experiment at 
85°C. The technetium concentration initially increased 33% 
from the initial concentration of 1.5 x lo^ ppb. The 
concentration then decreased and was essentially the same as 
the initial concentration by time the experiment ended. 
Despite the scatter in the data, the overall trend in the 
technetium concentration was such that it appears to not 
change much. Therefore, raising the temperature did not 
significantly increase the reaction kinetics. It is 
unlikely that the copper reduced the technetium to TC3O4. 
The technetium concentration in the experiments did not drop 
significantly, as the codes indicated it should (Figure 37) . 

The reason for the amount of scatter in the 
concentrations and the initial high peak is unknown, but it 
is unlikely caused by the release of the technetium retained 






144 
in the plumbing downstream of the reaction column (refer to 
the Blank-run Experiment in Section 4.2). This is because 
the maximum technetium concentration found in the Blank-run 
Experiment, 14 ppb, is far less than the increase of 490 ppb 
in the Copper Shot Experiment at 85°C. 

4.4.3 Copper Svunmary and Conclusions 

The results from the four copper experiments were 
consistent in that the technetium concentrations remained 
essentially unchanged. Despite the uncertainties with the 
FT-4C, the results from the three FT-4C copper experiments 
are considered valid because the effect of the technetium 
retention by the system is small compared to the starting 
and ending technetium concentrations of approximately 
1300 ppb to 1400 ppb. However, the codes indicated that the 
technetium concentration should be reduced to 
1.9 X lo'-^ ppb, which clearly does not agree with the 
experiments. The technetium concentration decreased 
approximately 200 ppb and then returned to the starting 
concentration of 1300 ppb in the Copper Shot Experiment at 
24 °C. The ending technetium concentrations were also 
essentially the same as the starting values (1200 ppb and 
14 00 ppb, respectively) in the DIW Copper Shot and the 
Copper Shot at 85 °C experiments, although there was much 
scatter. None of the experiments had the technetium 
concentration drop to the values indicated by the codes. 



145 
This disparity between the calculated and experimental 
results is most likely because a passivation layer, such as 
a copper oxide, developed on the metal's surface. This 
prevented any further reduction in technetium concentration 
beyond the initial drop, allowed the technetium 
concentrations in the effluent to return to their initial 
values, and may have allowed some of the technetium 
initially precipitated to redissolve. 

The passivation layer that formed was most likely CUjO. 
This was indicated by a greenish hue, indicative of CujO, 
that was on the copper surface after the experiment. As 
cuprite was forming, the copper was less capable of reducing 
the incoming oxidizing solution until eventually the 
passivation layer covered the entire surface, and the metal 
was prevented from further oxidation. Once the passivation 
layer formed, some of the technetium that was originally 
retained by the copper dissolved in the fresh incoming 
solution because technetium is very soluble in an oxidizing 
environment. 

Other possibilities for the disparity between the 
calculated and experimental results include technetium 
plate-out on the copper (which is a variation of the above 
explanation) and slow reaction kinetics; i.e., the time it 
would take for the reactant to react with the solution is 
much longer than the time in which the two were in contact 
with each other. Slow reaction kinetics is considered 



• 1 



146 
unlikely because 1) a reaction occurred that allowed the 
technetium concentration to be reduced temporarily but 
quickly, 2) an increase in temperature did not cause the 
technetium concentration to drop BDL or even cause it to 
continue to drop, and 3) the general trend in the 
experiments at 24 °C and 85 °C was similar. 

The ending technetium concentration in the Copper 
Powder Experiment was similar to that of the starting 
values. This was expected from the latter simulations, but, 
at the time of the experiment, the earlier simulations 
indicated that the technetium concentration should be 
reduced to BDL. Because the experiment starting 
concentration was close to the calculated ending 
concentration, there was no disparity between the 
experimental and calculated results. 

Comparing the surface area of the Copper Shot and 
Copper Powder experiments indicates that the protective 
passivation layer developed at essentially the same rate, 
regardless of the smaller surface area of the copper shot. 
This apparent similarity in rate may be a coincidence 
though, because there were two variables changed, the 
surface area and the initial technetium concentration. For 
the Copper Shot Experiment, it took 2 h for the technetium 
concentration to drop to a minimum of 1.13 x lo^ ppb and 
then begin to rise again. In the Copper Powder Experiment, 
it took 3 h to drop to a minimum of 1.5 x lo'^ ppb. The 



.;> ■■ 



.•^^ ^ 



147 
difference in the flowrates from a maximum of 50 mL/day in 
the copper shot experiments to 160 mL/day in the copper 
powder experiment did not affect the time for the 
passivation layer to develop, even though there was less 
time for the solution to react with the copper with the 
higher flowrate. 

Comparing the efficacy of technetium reduction as a 
function of surface area between the Copper Shot and Copper 
Powder experiments is difficult because two variables were 
changed, the surface area and the initial technetium 
concentration. The FT-4C Copper Shot Experiment at 24 °C 
reduced the technetium concentration by 200 ppb, which is 
considerably more than the reduction from 1 x lO'-^ ppb to 
1.5 X 10"*^ ppb in the copper powder experiment. However, in 
both cases, not all the technetium was removed from 
solution, regardless of surface area available. Once the 
passivation layer developed on the copper surface, further 
reaction of the copper and solution was prevented, halting 
further reduction in the technetium concentration. 

The results of these experiments do not show promise 
for copper reducing the rate of technetium migration in a 
geologic repository. 

4.5 Stannous Chloride Experiments 

Two stannous chloride experiments were conducted. The 
Beaker No-flow System, Figure lie, was used at ambient 



HK,- 



148 
conditions to study the effect of dissolved stannous 
chloride on '^Tc concentration reduction. This system was 
used because stannous chloride is a soluble salt. By using 
the Beaker System, the dissolved stannous chloride's redox 
effectiveness could be studied because the dissolved salt 
would remain in the system. 

A second stannous chloride experiment used the Bench- 
top Flow-through System shown in Figure lib. This Bench-top 
Stannous Chloride and Tuff Experiment was conducted to 
determine if the tuff would retain the precipitated 
technetium. The results from the Beaker and Bench-top 
experiments are discussed in Sections 4.5.1 and 4.5.2, 
respectively. 

4.5.1 Beaker Stannous Chloride Experiment 

The Beaker Stannous Chloride Experiment had the ^Tc- 
spiked solution added to a beaker containing dissolved 
stannous chloride. The technetium concentration decreased 
immediately from the starting concentration of 2 6 ppb to BDL 
(0.3 ppb) in the Beaker Stannous Chloride Experiment, 
Figure 38 Curve 1. Also, a white precipitate formed almost 
immediately after the technetium-spiked solution was mixed 
into the beaker containing the dissolved stannous chloride. 

A total of four samples was collected. The first was 
the starting ''Tc-spiked solution, the other three were from 
the beaker. The first sample had a technetium concentration 



149 




o\ 



\ 



\ 



i 



c 
E 






o 



CO 



B 
o 

I 

<u 
m 



\ 



'— »te«(** C^W( 



S:*^^-04«^^A^;t^««a^-L_ 



o 
o 



(qdd) U0UBJJU80U03 Lunijaugoai 



c 
« 
E 
•c 
a> 

Ol 
X 
LU 
« 



o 

(0 

3 
O 

c 
c 

(0 

Q. 
O 



o 
c 
« 
m 

? 

CO 

b. 
0) 

,^ 

(0 

m 



CO 
CO 

£ 

3 

.9 



!> 150 

of 26 ppb. The second and third samples had technetium 
concentrations that were BDL. The third sample was filtered 
(20 A pore size) to remove any precipitate before being 
analyzed for technetium concentration. The filter for the 
fourth sample was primed as per the procedure in Section 
3.4.3.2 to ensure that the technetium was not absorbed by 
the filter in the third sample. The remainder of the 
solution from the fourth sample was filtered and analyzed 
after priming the filter. Once again, the technetium 
concentration was BDL. 

Technetium has been observed to sorb onto glass. '^^ To 
determine if any technetium may have passed through the 
filter but sorbed onto the glass centrifuge tube, the tube 
was emptied and flushed with 0.1 M HNO3. The nitric acid 
flush was analyzed for "tc retention. Again, the ''Tc 
concentration was BDL. 

The white precipitate that formed when the technetium 
solution was added to the stannous chloride solution was 
most likely TC3O4. This is supported by the EQ3/6 
simulation of this experiment as shown in Figure 39. The 
model indicates that the technetium concentration (Curve 2) 
falls BDL at a reaction progress of 0.0019. This 
corresponds to the same point, as TC3O4 (Curve 3) begins to 
precipitate. In the experiment, the technetium 
concentration was BDL with the first sample collected, at 
which time the white precipitate had formed. Although the 



«, - 



•^ i.' ■ J-- * V 



O 
O 

d 



151 



^1 

a) 

Q 



8 



^ 
^ 



p 



c 
o 

00 



CM 

8 

d 



1 
Li 
a> 



0) 

O 

I 
<u 
m 



^3 





u. 




o 


■p 




3 


r 


fl 


f- 


o 


<u 




O 

E 


O 








f) 


<u 




UJ 







-: ■{ 



A 







8 


C 
00 


CM 


w 




c 




o 




^ 




•s 




8 


12 


o 



s 



I 



o 

o 



X 

UJ 
CN 
O 

c 
w 



m o 



Q. 

a. 



in 
o 



o 
o 



w 

>s 
(0 

2, 

0) 

E 
i- 



g. 
o 

■<r 
O 



c 


c 






!ii 


O 


ro 


F 




-1 


Q. 




(1 


<L 


<1l 








n 


(> 


CM 


(U 


I 


(O 


O 


(*5 


o 


o 



t 



d 



o 
in 
d 



O CM 

+ 



(s8|ouj) s|BJ9U!|/\| pajBjidioajd '(qdd) uojieJiuaouoQ Lunijauijoai 



in 

CM 

d 



8 

d 



o 

I- 



o 

3 

E 
w 

(O 

CO 

a 

HI 

T3 
CQ 

C 

a> 

E 
■c 

CL . 

UJ "^ 
2. % 

8.9- 
c o 
TO gj 

W Q- 

^2 

CO <u 
0) c 

m| 

°I 
o Q- 

CO S. 
Q. O 

E ^ 

o £ 
O Q. 



CO 

£ 

3 

at 



152 

■ ! .t ft-* ,:■ 

codes also indicate that the compound could be Tc(0H)2 
(Curve 4) , it does not appear to be this compound. A 
similar white precipitate was observed in the iron 
experiments, and the codes did not indicate that Tc(0H)2^ 
would form in those simulations. Another technetium 
compound that is known to exist, TCO2 (sometimes referred to 
as TC02-2H20 or Tc(0H)4), has been reported to be brown^° or a 
black sludge. ^^ This was not the compound, however, since 
the precipitate was clearly white with no signs of black or 
brown. The experiment is consistent with the results 
indicated by the codes. Because the detection limit for 
''tc is high, the only possible conclusion is that the 
stannous chloride reduced the technetium concentration from 
2 6 ppb to BDL (0.3 ppb) . 

4.5.2 Bench-Top Stannous Chloride Experiment 

The Bench-top Stannous Chloride Experiment used a 
reaction column loaded with 1.71 g each of stannous chloride 
and crushed tuff. It was conducted to determine if tuff in 
the reaction column would retain the precipitated 
technetium. Because the Beaker Stannous Chloride Experiment 
had a ''tc concentration BDL, '^'"Tc was used because of its 
8 orders of magnitude lower detection limit (Section 3.1). 



JThere is no recorded data on the color of this compound. 



153 
A 9.4 X 10"^ ppb '^^Tc-spiked solution was pumped at a 
flowrate of 152 mL/day through the column at 24 °C. 

The results from this Bench-top Stannous Chloride 
Experiment are shown in Figure 38 Curve 2. As shown in 
Curve 2, the technetium concentration in the effluent 
decreased from the starting concentration of 9.4 x lO"^ ppb 
to 6.5 X 10"*^ ppb within the first 6 h of the experiment. 
These values are above the detectable limits for '^""Tc for 
this experiment (detection limit was 7.6 x lo'^ PPb) • This 
initial decrease in technetium concentration was followed by 
an increase with an ending concentration of 1.3 x lo'-' ppb 
after 8 days, which is slightly over the starting value. 

The first sample collected contained a white 
precipitate. The total technetium activity of this entire 
1.1 -q sample was 57 Bq before it was filtered (0.45-/im 
openings), and 9.83 Bq after it was filtered. Very little 
precipitate (0.38 g) remained in the original sample vial 
and was therefore not filtered. This is being referred to 
as the residue. The residue was analyzed and had a '^""Tc 
activity of 4.91 Bq. If it is assumed that the residue was 
all precipitate, then: 

(1 - ^'l^^^ Bq )x 100 = 83% 
57 



154 
indicating that 83% of the technetium that was detected in 
the first sample was in the form of a precipitate. Only 17% 
was still in solution. 

The precipitate observed in the first sample was most 
likely TC3O4, caused by the reduction of TCO4" with stannous 
chloride. This is supported by the comparison with the 
computer simulation, which shows that the measured 
technetium concentration decreased below the concentration 
at which EQ3/6 predicts that TC3O4 should start to 
precipitate (Figure 40) . This is represented by following 
Curve 5, which (1) connects the point where the measured 
technetium concentration was minimum (Curve 1), (2) equals 
that of the calculated technetium concentration (Curve 2) ; 
and (3) follows Curve 6 vertically down to the point where 
it intersects the moles of TC3O4 precipitated (Curve 3) . 
The technetium concentration decreases substantially just as 
the TC3O4 begins to precipitate. 

Although the technetium concentration never went BDL in 
the experiment, it dropped over 2 orders of magnitude. It 
is highly unlikely that the technetium was reduced to any 
form with lower oxidation states than TC3O4 because the 
concentration never went BDL, as it should if lower 
oxidation states were obtained. It is most likely that the 
technetium concentration did not drop BDL because the 
stannous chloride dissolved too quickly to reduce all the 
TC3O4. It is likely that the solubility-controlling 

■'' * ^ """ ■'<■ « .:■ -j'f r f.. --^ 

(■ ■ '■■■•■ 




CO 

E 



(s9|oai) S|Bjau!i/\| pajendpajd '(qdd) uoijeJiuaouoQ 



■D 
0) 



■o 



c 

(0 
(A 

c 
o 



0) 

o 

c 

8 



c 

o 
a> 

I- 



E 
(0 

eo 
O 

liJ 



c 

CO 

c 

E 

Q. 

o 
■o 



o 

*— c/l 

CD C 

2 TO 

I i_ 

■C 03 

O C 

CD ^ 

C CO 

Q. Q. 

E o 

o 2^ 

O Q. 



.s? 



155 



156 
compound was TC3O4 because the precipitate was white, 
similar to those observed in the iron experiment and the 
Beaker Stannous Chloride Experiment. The technetium 
concentration was well above the 1.8 x 10"' ppb (reaction 
progress of 0.0032) concentration where the codes predict 
Tc(0H)2 will precipitate. 

The computer simulations in Figure 39 and Figure 40 
look different because the curves are plotted as straight 
lines between calculated data points. These points were 
calculated at different values of reaction progress in the 
two different simulations. The main information to be 
obtained from these figures is that the codes indicated that 
the technetium concentration should drop substantially BDL, 
and the first technetium compound to precipitate was TC3O4. 

4.5.3 Stannous Chloride Summary and Conclusions 

The ''tc concentration in the Beaker Experiment dropped 
to BDL (0.3 ppb) and stayed there for the duration of the 
experiment (2 days) . Using '^""Tc to improve the detection 
limit, concentrations in the effluent samples from the 
Bench-top Experiment dropped to a minimum of 6.5 x 10"* Ppb. 
This concentration was well above detectable limits for 
'^""Tc, but it was still 5 orders of magnitude greater than 
the calculated results from the computer codes. Following 
the minimum, the technetium concentration immediately 
(within 1 day) increased back toward the starting 



' 157 

concentration and eventually exceeded it slightly by the 
time the experiment was terminated on day 8. 

The technetium concentration decreased and then 
returned to that of the entering solution because the 
stannous chloride dissolved and was expelled from the 
column. Initially, the stannous chloride dissolved when the 
'^"'Tc-spiked J-13 solution encountered the stannous chloride 
in the reaction column. The solution became reducing, 
causing the technetium in the solution to precipitate. As 
the solution continued to enter the reaction column, the 
concentration of the dissolved stannous chloride decreased, 
and eventually the stannous chloride concentration was not 
high enough to yield a reducing solution environment. Any 
technetium that precipitated and/or was retained by the tuff 
then dissolved because the solution environment was now 
oxidizing because of the oxidizing solution entering the 
column. This then caused the technetium concentration in 
the effluent samples to be greater than that of the starting 
concentration. 

Comparing the results of the experiments with those 
indicated by the codes shows that since the stannous 
chloride concentration was constant in the Beaker 
Experiment, the "tc concentration remained BDL as expected. 
However, the flow-through experimental results were not in 
agreement with the results indicated by the codes because 
the stannous chloride dissolved and exited the column. 



^ 158 

Stannous chloride does not show promise for reducing 
the rate of technetium migration at a geologic repository 
based on the results of the Bench-top Stannous Chloride 
Experiment. These results show that stannous chloride 
reduced the technetium concentration only for the limited 
time that it remained in the column before it dissolved and 
was expelled from the column. The white precipitate that 
exited the reaction column also indicates that even though 
the stannous chloride temporarily reduced the technetium 
concentration, the precipitate was transported through the 
crushed tuff in the column. It is unsure how fast this 
mobile precipitate would travel because it could get stuck 
in the corners of fragmented tuff, or it may preferentially 
travel along the fault lines of the repository. This has 
significant ramifications because it may not be enough to 
only chemically reduce the technetium; the technetium must 
also be immobilized. Otherwise, the mobile technetium 
precipitate, albeit chemically reduced, could be oxidized 
once it encountered an oxidizing environment. Once it is 
oxidized and redissolved, it would likely have greater 
mobility than it would as a finely divided precipitate. 

Using stannous chloride as a reducing agent also poses 
the possible disadvantage of increasing the degradation of 
the spent fuel waste package because of the acidic 
environment it may create. The advantage of using stannous 
chloride, namely its faster reaction kinetics, does not 



• - 159 

outweigh the disadvantage that the stannous chloride rapidly 
dissolves. Nor does it outweigh the disadvantage that the 
precipitate that formed was mobile in the crushed tuff of 
this experiment. Because of its high solubility, all the 
stannous chloride in a waste package could potentially be 
exhausted during the first groundwater infiltration. Should 
there be more than one occurrence, there would not be a 
reactant available to reduce the incoming groundwater and 
inhibit technetium migration away from the repository. 

4.6 FT-4C Tin Shot Experiments 

Two experiments pumped "Tc-spiked J-13 through 
reaction columns containing tin shot. One was at room 
temperature, 24 °C, and the other was at an elevated 
temperature, 85 °C. Both of the Tin Shot experiments used 
the FT-4C, Figure 11a. It is known that the FT-4C system 
had problems with technetium being retained in the plumbing 
and later being released in the same or subsequent 
experiment, as discussed in the Blank-run Experiment, 
Section 4.2. However, the concentrations of technetium in 
the Tin Shot experiments were substantially higher than was 
found in the Blank-run Experiment. Therefore, this 
retention phenomena had no significant impact on the tin 
data, and the results from these experiments are credible. 






>--Mv ,. . ■ ■-' 160 

4.6.1 Tin Shot Experiment at 2 4 °C 

The technetium concentrations versus time from the 
FT-4C ''tc Tin Shot Experiment at 24 °C are represented by 
Figure 41 Curve 1. The technetium concentration in this 
experiment had a maximum drop from the starting 
concentration to approximately 2 ppb within the first 3 h of 
the experiment. This drop lasted less than 4 h, and the 
technetium concentration then increased to 972 ppb (3% below 
the starting concentration) , which is within the range of 
uncertainty (4.9%). The technetium concentration then 
fluctuated within 14% of the starting concentration, having 
a final technetium concentration of 900 ppb, 9% below the 
starting concentration. 

Figure 42 shows the comparison of the tin shot 
experiments and the EQ3/6 simulation of these experiments. 
The codes predicted that the technetium concentration would 
drop to 1 X 10-12 ppj^ gg ^^^Q^ ^^^ Tc(0H)2 would precipitate 
out. The experiments never had the technetium drop BDL (as 
it would have if it dropped to the levels indicated by the 
codes) . 

4.6.2 Tin Shot Experiment at 85 °C 

The technetium concentration versus time from the FT-4C 

99m 

Tc Tin Shot Experiment at 85 °C is represented by Figure 41 
Curve 2. This experiment also had an initial drop in 
technetium concentration. However, it was only an 8.5% drop 



161 



i* 



^ 



\ 




o 

o 

m 

00 

(0 

® fsT 

X O 

III --' 



V) 

c 

F 



o 

o 

■««■ 

CM 

CO 

c 

^p 

S3 ^ 

Ql 3 

X O 

liJ ^-^ 



CO 

F 



TO 

a> 

E 



P 



^ 



8 8 

CO t^ 



o 
o 



o 
o 
in 



§ 



o 
o 

CO 



(qdd) U0UBJJU80U03 Lunijaugoai 



t 






162 



s 

« 



u 



c 


^^ 




CM 


fc 




(11 


£ 


n 




X 


() 


111 








o 




■<- 





CO 



o 
O 



■3 <^ 



c 
o 

(G 



c ^ 




r 















(Q 












E 


iB 




w 


1 
















CO 








LLI 




1- 




1 


r 


« 




(O 

CO 

CD 

E 



o 
o 
o 



$ 



O 



■* 




CM 








m 








r 




<u 


' ' 


h 


(1) 


S 


3 


X 


() 


LLI 




C 




1- 






(S9|ouj) sjonpojd pajejidpajd >? (qdd) uo!jbj}U90uoo ainuauqoai 



3 

E 
CO 
CD 
CO 

o 

LU 

<u 

.c 

x> 

c 

CO 

m 
c 
d) 

E 

'i— 

(U 
Q. 
X 

LLI 

O 

w 

i- 
o 



0) 



c 
o 
« 

CO 
Q. 

E 
o 
O 



5! 
u. 



163 
from 1011 ppb to 92 5 ppb. The technetium concentration then 
increased 3% to 985 ppb. It continued to fluctuate within 
12% of the starting concentration, reaching a minimum of 
856 ppb 3 days after the start of the experiment. The 
technetium concentration then increased, and 5 days after 
the experiment started, it was 102 ppb, which is within the 
uncertainty limits (5%) of the starting concentration. The 
technetium concentration essentially remained equal to the 
starting concentration for the remainder of the experiment. 

Because the reaction kinetics are faster at higher 
temperatures, it is possible that this experiment did not 
quite agree with that of the 24°C experiment. The initial 
drop in technetium concentration was much less dramatic in 
this experiment than that of the Tin Shot Experiment at 
24 °C, which is an example of the effect of faster reaction 
kinetics at higher temperatures. It is possible that the 
technetium concentration dropped as much as in the 
experiment at 24 °C, but the drop may have been missed by the 
time the first sample was collected (50 min after the start 
of the experiment) because of the faster reaction kinetics. 

The tin shot experiments did not agree with the EQ3/6 
simulation as the experiments never had the technetium drop 
BDL (Figure 42) . The tin shot had essentially no effect on 
reducing the concentration of technetium in solution. 

The "tc concentration in the tin experiments was 
approximately 1000 ppb compared with a maximum of 14 ppb in 



164 
the Blank-run Experiment, Section 4.2. Therefore, the 
discrepancy between the results from the experiments and the 
computer simulation is not due to the system retaining 
technetium (in the plumbing downstream from the reaction 
column) . ^:5 i. ' w ! a '.':'■'* ^ 

- .. ■ ■'■ ;^ '■ > ' ■ J! ^ t I '' :■ 
4.6.3 Tin Summary and Conclusions 

Similar to the results for copper, it is believed that 
a passivation layer formed on the tin that prevented the tin 
from reducing the technetium concentration after the initial 
drop. This is supported by the fact that tin forms an 
insoluble film of SnO in waters with a pH of 7.4 and 8.6 (J- 
13 is within this range) .^° The passivation layer that 
developed on the tin in the experiments (1) prevented 
further reduction of the solution environment, (2) allowed 
the solution that entered the reaction column to remain in 
its original oxidizing state, and (3) allowed the oxidizing 
solution to dissolve any technetium that was previously 
reduced. Evidence for the third statement is that the 
technetium concentration would be expected to increase above 
the starting concentration after the initial drop, which it 
did. 

The EQ3/6 simulations indicated that cassiterite (Sn02) 
would form and precipitate, but not romarchite (SnO) . 
According to known tin reactions, the formation of 
romarchite would be necessary to provide a protective 



165 
passivation layer on the tin/° Therefore, in an attempt to 
have romarchite precipitate in the EQ3/6 simulations, 
cassiterite was suppressed. Neither romarchite nor any 
other compound containing tin was indicated to precipitate. 
As the reaction path progressed, all the added tin went into 
solution. Another simulation had romarchite added as the 
reactant, thus forcing romarchite into the system 
(cassiterite was suppressed again) . The simulation for that 
run also indicated that all the tin would be in solution 
with no tin compounds precipitating. Even though the codes 
do not predict that SnO forms, this may be in error because 
experimental evidence^° suggests that it actually does. 

The possibility of slow reaction kinetics being 
responsible for the discrepancy in technetium concentrations 
between the experiments and the codes is unlikely because 
the technetium concentration was quickly reduced 3 orders of 
magnitude in the 24 °C experiment. In the 85 °C experiment, 
the drop was not as substantial, but it did drop. Thus, the 
initial reaction was quite rapid, but formation of the 
passivation layer prevented further reaction. 

The results of these experiments do not show promise 
for tin reducing the rate of technetium migration in a 
geologic repository. 

,■■„■■ ■ ■ ■■ i <■ , " ■ ■'■'' > t 
'*■ •■ .. ^ * 

r. »' -«■*, ■ *-. '■'..; ■ * .. ■ 'V ■--" 

i 






! ^ , . ^ . .' ' ^ ■ ■' I f\ 



CHAPTER 5 
SUMMARY AND CONCLUSIONS 

The conclusions listed here are based upon computer 
simulations performed at the University of Florida and 
during and after interning at PNNL, and on the experiments 
conducted at PNNL. This work was performed to evaluate 
different reactants with a potential for reducing the rate 
of ''tc migration under the scenario of waste cask breach 
and groundwater infiltration at a geologic repository. 
Reducing the rate of technetium migration at a geologic 
repository is of interest because ''tc is expected to be 
present as pertechnetate (TcO^') , which is highly soluble 
under oxidizing conditions. It has a very long half -life of 
213,000 years. ^'2° However, technetium is less soluble and 
may be more strongly adsorbed on minerals when it is reduced 
to lower valance-state compounds (TCO2, TC3O4, Tc(0H)2, 
Tc(OH) , Tc, etc.) .2.11.20.21 

Therefore, iron, copper, stannous chloride, and tin 
were investigated in the hope that they would lower the 
redox potential of the solution environment and thus 
precipitate technetium from solution. Because the reduced 
technetium compounds are much less soluble, they should 
remain close to where they precipitate. . 



166 



167 
Experimentally, iron was the only reducing agent 
investigated in these experiments that shows promise for 
reducing the rate of technetium migration from a geologic 
repository. The '^""Tc concentration decreased from a 
starting concentration of 9.4 x lo"^ ppb in the flow-through 
experiment to near detection limits (1 x lo"'' ppb) . Then it 
increased, reaching 3 x lo"^ ppb at the experiment's end 
(38 days) . In the static experiments, the '^""Tc 
concentration was reduced to BDL (1 x lo"'' ppb) . 

Comparison of the two static experiments (one at 70 °C 
and the other at 24 °C) showed that the increase in 
temperature substantially reduced the time it took to 
decrease the technetium concentration BDL. In the 70°C 
experiment, the technetium concentration dropped BDL within 
the first day, versus more than 13 days required in the 24 °C 
experiment. This was caused by the slow reaction kinetics 
at the lower temperature. 

In each of the iron powder experiments, the technetium 
concentrations were reduced for the entire time that the 
experiments were conducted; however, the longest experiment 
lasted only 56 days. Follow-up studies should be conducted 
to determine whether iron can be expected to reduce 
technetium concentrations on a geologic time scale. If so, 
adding iron into the waste package design for spent nuclear 
fuel should reduce the rate of technetium migration from a 
geologic repository should a breached waste package be 



168 
contacted by groundwater. The technetium would be less 
mobile in the reduced and precipitated state. 

Comparing the results of this research with those 
calculated by the EQ3/6 codes indicates that the measured 
technetium concentrations in the static iron powder 
experiments were consistent with those indicated by the 
codes, as both experiments terminated with technetium 
concentrations BDL. The pH measurements were also in 
agreement with the codes for these two experiments. 
However, the measured minimum Eh value (-0.28 mV) was 
substantially higher than the minimum indicated by the codes 
(-492 mV) for the 70°C experiment. The measured minimum Eh 
value (-343 mV) for the 24 °C experiment was also 
substantially higher than the minimum indicated by the codes 
(-444 mV) . This was because the experimental environment 
was not as anoxic as hoped. Although the experiments were 
conducted in sealed vessels, there was oxygen in-leakage. 

The iron powder flow-through experiments were not 
modeled correctly by the codes for three reasons. First, 
the reaction kinetics of iron are slow, and therefore the 
values calculated by the codes were never realized in the 
limited time frame that the iron and technetium were in 
contact with each other (approximately 1.5 h) . Second, the 
EQ3/6 codes cannot accurately model a solid-centered flow- 
through open system as in the experimental flow-through 






*. 



169 
systems."^ Finally, the EQ3/6 codes cannot model sorption. 
In the Bench-top '^^"^c Iron Powder Experiment, the technetium 
concentration decreased because it was being sorbed as well 
as reduced by the iron. 

The results from the four copper experiments were 
consistent in that the technetium concentration remained 
essentially unchanged. In two of the experiments, there was 
an initial drop in technetium concentration, after which the 
concentration returned to the starting values. This 
occurred because the technetium was first precipitated, and 
then it redissolved once the CU2O passivation layer 
developed and prevented the metal from further oxidation. 
These experimental results are in contrast to the 
calculations which show that the technetium concentration 
should have dropped to 2 x lo"^ ppb. The fact that the 
technetium concentration remained essentially unchanged for 
the copper experiments indicates that copper would not prove 
promising for reducing the rate of technetium migration at a 
geologic repository. 

When stannous chloride was used in a static experiment, 
the technetium in solution precipitated instantaneously. 
This was consistent with the results indicated by the 



v W 



The EQ3/6 codes are continually being updated. It is 
planned to revise the code so it will be capable of modeling 
a solid-centered flow-through open system as a regular 
feature in the future. ^^ After adding this feature, the 
codes would be capable of modeling the experiments. 



computer codes. However, in the flow-through experiment, 
the technetium concentrations returned to, or were higher 
than, the starting concentration once the stannous chloride 
dissolved and was expelled from the reaction column. The 
technetium precipitate was also mobile in the crushed tuff 
of the flow-through laboratory experiment. Therefore, it 
might also be mobile in a repository environment. It is 
unsure how fast this mobile precipitate would travel because 
it could get stuck in the corners of fragmented tuff, or it 
could preferentially travel along the fault lines of the 
repository. Even if the technetium precipitate moved guite 
slowly, it might eventually reach an oxidizing environment 
where it would redissolve and then migrate along with the 
water. 

The tin shot experiments at 85 °C and 24 °C were 
consistent in that the technetium concentration remained 
essentially unchanged in both experiments. There was an 
initial drop in the technetium concentration, after which it 
returned to near the starting values. This was most likely 
because of the formation of a protective passivation layer 
of SnO on the tin. As a result, the experimental results 
are not in agreement with those indicated by the codes 
because the codes cannot model a passivation layer. 
Although the codes cannot model this type of flow-through 
experiment, it is also unlikely that a stagnant experiment 



171 
would be modeled by the codes correctly. The passivation 
layer would likely develop in that experiment also. 

The possibility that slow reaction kinetics in the tin 
shot experiments was responsible for the discrepancy in 
technetium concentrations between the experiments and the 
codes is unlikely because the technetium concentration was 
quickly reduced 3 orders of magnitude in the 24 °C 
experiment. In the 85 °C experiment, the drop was not as 
substantial, but it did drop. Thus, the initial reaction 
was quite rapid, but formation of the passivation layer 
prevented further reaction. 

Iron appears promising for reducing the rate of 
technetium migration from a geologic repository. Copper, 
tin, and stannous chloride did not appear promising. 






'r^^-^'iA 






CHAPTER 6 
, RECOMMENDATIONS 

A flow-through experiment with iron showed that the 
'^"^c concentration dropped from a starting value of 
9.4 X 10"^ ppb to approximately 1 x lO'^ ppb in under 2 h. 
The technetium concentration remained near this detection 
limit for 13 days and then began to increase, reaching a 
maximum concentration of 3 x lo'^ ppb after 3 days. It is 
believed that the rise in technetium concentration in the 
latter half of the experiment was caused by the sorption of 
technetium by the iron, thereby passivating its surface. 
Future experiments lasting substantially longer times would 
be beneficial to determine if the technetium concentration 
would remain at reduced values or continue to increase as 
passivation of the iron surface became more effective. 
Analyses of the iron at the end of the experiments would 
also be beneficial to determine if the technetium was 
actually being sorbed by the iron. 

The two Sealed Vessel experiments showed that the iron 
reduced the '^'"Tc concentration from 2.3 x lo"^ ppb to BDL, 
i.e., less than 1 x lO"'' ppb. Although these experiments 
were conducted in a sealed vessel, there was oxygen in- 
leakage. Therefore, the Eh values were much greater than 



172 



173 
indicated by the codes. Experiments using a system similar 
to the Sealed Vessel System could be performed in an oxygen- 
free environment. This could be obtained by using a 
specially designed leak-free system in an inert-gas-purged 
glovebox. These experiments would alleviate the problem of 
oxygen infiltration that occurred in the Sealed Vessel 
experiments. These experiments could then have a known 
concentration of oxygen added to the system to determine the 
effect of oxygen on the reduction of technetium. Because 
the proposed geologic repository at Yucca Mountain would not 
be in an anoxic environment, understanding the rate of 
technetium concentration reduction as a function of oxygen 
concentration would be beneficial. Although the technetium 
concentration was reduced BDL for the entire 15 and 56 days 
that the experiments were conducted and would not be 
expected to increase in longer experiments, the new 
experiments could be conducted for substantially longer 
times to ascertain that it would not increase. Analyses 
could also be conducted on the precipitates and the iron 
after the experiments are terminated to determine the 
oxidation states of the reduced technetium and oxidized 
iron. 

To take advantage of the 8 orders of magnitude 
difference in detection levels, a static experiment using 

""Tc instead of ''tc is recommended to gain further 
information on the reducing capabilities of stannous 



174 
1 - 

chloride on technetium. In the static experiment using 
"tc, the concentration was BDL as the spiked solution was 
added to a dissolved stannous chloride solution. Using '^"Vc 
instead of ''tc would allow the lowest technetium 
concentration to be determined and possibly determine the 
solubility-controlling technetium compound. The precipitate 
that formed as the spiked solution was added to the stannous 
chloride solution can be analyzed to determine its chemical 
form. In similar flow-through experiments, the reaction 
column could also be analyzed to determine if any of the 
technetium gets caught on the corners of the fragmented 
tuff, or if it travels with the fluid flow. A similar 
experiment using uncrushed tuff could be conducted to 
determine the mobility of the precipitate in this medium. 

The copper powder experiment used a technetium 
concentration of 9.4 x lO'"^ ppb, which was close to the 
minimum indicated by the codes, 3.4 x lo"^ ppb. Thus the 
technetium concentration remained essentially unchanged, 
except for a temporary initial drop. An experiment should 
be conducted using a starting technetium concentration much 
greater than the minimum predicted by the codes. Then it 
could be determined if the technetium concentration would 
drop to the predicted values or remain essentially unchanged 
as it did in the experiment described here. 

To obtain these much greater starting concentrations, 
these new experiments may be conducted with a combination of 



175 



''Tc and '^""Tc. They may also be conducted with another 



tracer material that should chemically behave as technetium, 
e.g., rhenium (Re). Rhenium has radioactive and non- 
radioactive isotopes. Experiments using large 
concentrations of rhenium can be conducted using the non- 
radioactive isotope. The radioactive isotope could then be 
used for experiments with trace concentrations. Therefore, 
samples from experiments that use high concentrations of the 
non-radioactive isotopes can be analyzed chemically. 
Samples from experiments that use trace amounts of the 
radioactive isotopes can be analyzed radiochemically as in 
the technetium experiments. Rhenium-183 is a gamma emitter 
with a relatively short half-life of 70 days," and low 
detection limits similar to '^""Tc can be obtained (see 
Table 3) . 

The disadvantages versus advantages would need to be 
researched before conducting any experiments. Using ^^^Re 
may have a disadvantage because of the complex redox 
reactions associated with this isotope. The advantage of 
using large concentrations without radioactivity may 
outweigh those disadvantages due to the rules, regulations, 
and costs of conducting radioactive experiments. The 
financial feasibility of conducting these experiments also 
needs to be considered. Conducting chemical analyses may 
cost significantly more than conducting radiochemical 
analyses, especially if the chemical analyses must be 



176 
contracted out. However, one could use high concentrations 
of the non-radioactive rhenium and spike it with ^^^Re to 
follow what occurs radioactively . Obtaining ^^^Re may also 
be financially prohibitive. It would also need to be 
ascertained that rhenium chemically behaves as technetium. 
Otherwise, the experimental results would be useless for 
comparison to those of the technetium experiments, or for 
application to reducing the rate of technetium migration at 
a geologic repository. 






CHAPTER 7 
EPILOGUE 

During this investigation, the author received 
inquiries regarding the possibility of using the information 
gained through this research to prevent the migration of 
technetium in the ground at sites that contain organic 
contaminants (especially trichlorethylene, TCE) along with 
technetium. The inquiring groups had already ascertained 
that iron would bind the organic contaminants and that the 
site in question was oxidizing. The information obtained 
from this dissertation research suggests that the use of 
iron would reduce the oxidation state of technetium, thus 
causing it to precipitate. The rate of technetium migration 
would be expected to also be much less as a solid 
precipitate than as an anion (Tc04') in solution. Thus it 
appears that the addition of iron to the contaminated sites 
would bind both forms of contaminants. The soil and 
associated contaminants could then be dug up and properly 
disposed of as hazardous mixed waste. 

PNNL is currently studying In Situ Redox Manipulation 
(ISRM) , originally called Redox Manipulation, to cheaply 
stop the flow of highly toxic chromium toward the nearby 
Columbia River." ISRM involves creating an iron (II) zone 



177 •' T L * 






178 
by injecting a chemical reagent (containing sodium 
dithionite and potassium carbonate/bicarbonate pH buffers) 
into the aquifer through a groundwater well. This chemical 
causes the naturally occurring iron (III) in the aquifer 
sediments to be reduced to iron (II) . The reduced iron is 
then capable of reducing the contaminants in the ground that 
are redox-sensitive. Technetium is among the list of redox 
sensitive contaminants that PNNL wants to remove from its 
site. This is very similar to the concept of placing iron 
in a geologic repository to reduce the oxidation state of 
technetium and hence reduce its concentration should a 
breach occur in a waste package and be infiltrated with 
groundwater. The difference is that PNNL uses the iron 
present in the soil rather than adding it to the waste 
environment. PNNL's technology is being hailed as a means 
to treat contaminated groundwater throughout the world. 

The kinetic and thermodynamic properties of technetium 
are still being researched. There are many unknowns and 
uncertainties in these data because of the complex nature of 
technetium's chemical properties. One such investigation is 
being conducted at Los Alamos National Laboratory to study 
technetium attenuation and the role of mineral interactions 
with technetium. This study will investigate the mechanisms 
of surface-related sorption of technetium on surfaces 
containing iron(Il).5^ 



GLOSSARY 



Abscissa 



The horizontal coordinate of a point 
in a plane Cartesian coordinate system 
obtained by measuring parallel to the 
X-axis. 



Activity 



The value that replaces concentration 
in a thermodynamically correct 
equilibrium expression. Also, a 
measure of radioactive decay in units 
of Bq or Ci. 



Activity 
coefficient 



The number by which the concentration 
must be multiplied to give the 
activity. 



Actinide 



Elements that are in the seventh 
period of the periodic table. Uranium 
and Plutonium are members of the 
actinide series. 



Adsorb 



The process by which a substance 
becomes attached to the surface of 
another substance. 



Anoxic 
BDL 



Without oxygen. 

Below detectable limits, 



Chemical 
equilibrium 



■ ■■^. ^ 



Conduction band 



The state in which the forward and 
reverse rates of all reactions are 
equal so that the concentrations of 
all species remain constant. 

In a crystal, the band contains 
electrons that are not bound and 
migrate throughout the crystal. 



179 



180 



Debye-Hiickel 
equation 



Gives the activity coefficient, y, as 
a function of ionic strength, J. The 
extended Debye-Hiickel equation, 
applicable to ionic strengths to about 
0.1 M, is 

log Y = [-0.51z^VI]/[l + (aVJ/305)] 
where z is the ionic charge and a is 
the effective hydrated radius in 
picometers. 



Decay heat 
Disequilibrium 



Fluid-centered 



Florescence 



Heat generated by nuclear decay of 
isotopes. 

The state in which the forward and 
reverse rates of all reactions are not 
equal so that the concentrations of 
all species are changing. 

The view as would be seen by a packet 
of fluid traveling through a medium. 

The property of emitting radiation as 
the result of absorption of radiation 
from some other source. The emitted 
radiation persists only as long as the 
exposure is subjected to the radiation 
that may be either electrified 
particles or waves. The fluorescent 
radiation generally has a longer wave 
length than that of the absorbed 
radiation. 



Forbidden band 



Geochemical 



Energy band within a crystal where 
electrons are normally forbidden. 

The chemical and geological properties 
of a substance. 



Half-life 



Integral number 
of moles 



The time required for one half the 
atoms in a sample of a radioactive 
element to decay. 

The total moles added in the 
incremental steps taken by the EQ3/6 
codes. 



<» "■ ^ 



181 



Interstitial 
states 



Places for electrons in the normally 
forbidden band made by adding an 
impurity into the lattice. 



Isomer 



Nuclei with the same mass number, A, 
and the same number of protons, Z, but 
with different energies. 



Isotope 



An element having various values of 
the mass number A, for a given number 
of protons, Z. Isotopes behave 
similarly in chemical reactions, but 
their nuclear characteristics may be 
markedly different. 



Iterative 
technique 



Technique where an approximation for 
solving a numerical equation is made 
on the basis of a previous 
approximation. This processes is 
performed repeatedly to successively 
compute better and better 
approximations . 



Kinetics 



The rate of change in a chemical 
system. 



"m" in '5"^c 



Metastable. Nuclei with a given mass 
number. A, and number of protons, Z, 
that exists temporarily in a state 
having more energy than the ground 
state, corresponding to that of A and 



Z. See 



'isomer' 



Molality 



The number of moles of solute per 
kilogram of solvent. 



Ordinate 



The vertical coordinate of a point in 
a plane Cartesian coordinate system 
obtained by measuring parallel to the 
y-axis. 



Oxic 



Oxidation state 



Containing oxygen. 

A bookkeeping device used to tell how 
many electrons have been gained or 
lost by a neutral atom when it forms a 
compound. 



182 



Oxygen fugacity 



The activity of oxygen gas. A = P02Y02 



Passivating layer A layer on the compound that makes it 

inactive or less reactive. 



Pentavalent 

Pertechnetate 
PH 



pHCl 

Phase assemblage 

Porbaix diagram 

Precipitate 

Radionuclide 
Reactant 

Redox reaction 

Reducing agent 

Reduction 



Having five electrons in the outer 
shell. 

The chemical name for TCO4". 

Defined as pH = - log Ah+, where Ah+ is 
the activity of H*. In most 
approximate applications, the pH is 
taken as -log[H''] . 

Defined as pHCl = - log Ahci, where Ahci 
is the activity of HCl. 

The collection of chemical phases that 
can occur. 

A diagram of Eh versus pH illustrating 
the possible chemical species formed 
in a system. 

The solid substance that separates 
from a solution. 

Radioactive nuclide. 

The species that is consumed in a 
chemical reaction. It appears on the 
left side of a chemical equation. 

A chemical reaction involving the 
transfer of elections from one element 
to another. 

A substance that donates electrons in 
a chemical reaction. Also called 
Reductant. 

A gain in electrons or a lowering of 
the oxidation state. 



183 



Repository 



A site that may be designated for the 
long-term storage of radioactive spent 
fuel generated from nuclear reactors 
that is not at the reactor sites 
themselves. 



SCE 



Saturated Calomel Electrode is the 
common reference electrode based on 
the half-reaction Hg2Cl2(s) + 2e" = 
2Hg(l) + 2C1'. 



Solid-centered 



Solubility 



The view as would be seen by a solid 
as fluid travels through the medium. 

Is the mass of a substance, a solid or 
liquid, contained in a solution in 
equilibrium with an excess of the 
substance. 



Sorb (sorption) 



Supersaturated 
solution 



Thermocouple 



The method by which a chemical species 
is retained by another without a 
chemical reaction occurring. 

A solution that contains more 
dissolved solute than would be present 
at equilibrium. 

An electrical junction across which a 
temperature-dependent voltage exists. 
Theremocouples are calibrated to 
measure temperatures and usually 
consist of two dissimilar metals in 
contact with each other. 



Thermodynamic 



Titration 



Relating to a system of atoms, 
molecules, colloidal particles, or 
larger bodies considered as an 
isolated group in the study of 
thermodynamic processes. 

A procedure in which a substance 
(titrant) is carefully added to 
another (analyte) until a complete 
reaction has occurred. The quantity 
of titrant required for a complete 
reaction tells how much analyte is 
present. 



184 



Tuff 



Rock composed of fine volcanic 
rock/dust that has been fused together 
by heat. 



Vadose 



Water or solution above the 
groundwater table and located in the 
tuffaceous rock (see tuff) . 



Valance band 



The band that contains electrons that 
are bound to specific lattice sites 
within the crystal. 



Valence state 



Numerical value equal to the number of 
electrons in its outer shell. 



V ». 






I 



APPENDIX A 
THE EQ3/6 COMPUTER CODES 

Computational-analysis was conducted using the EQ3/6 
software package. ^^•^''•^^•^' The EQ3/6 codes are actually two 
codes that characterize the geochemical behavior of aqueous 
systems, EQ3NR and EQ6 ; a data file preprocessor, EQPT; a 
software library, EQLIB; and supporting thermodynamic data 
files. 

This software package was originally developed at 
Northwestern University in the 1970s "to model seawater- 
basalt interaction in mid-ocean ridge hydrothermal 
systems."^' It has been revised numerous times since then. 
It is currently the property of Lawrence Livermore National 
Laboratory (LLNL) . The EQ3/6 codes were written in 
Fortran 77 and were developed to run within the UNIX 
operating system on workstations and supercomputers. 

EQ3NR and EQ6 both require using one of the 
thermodynamic data files. Currently, there are five of 
these files that are used with the EQ3/6 codes. These files 
are all formatted ASCII files called DATAO . * where * is COM, 
SUP, NEA, HMW, and PIT. However, EQ3NR and EQ6 require 
using unformatted data files. EQPT is the program used to 
convert the formatted data files into the required 

185 



186 
corresponding unformatted data files called DATAl . * files. 
Each of the five data files contains the activity- 
coefficient parameters and standard-state thermodynamic 
data. These data are needed to determine the activity 
coefficients. The files correspond to the equations used 
for modeling the activity coefficients of the aqueous 
species it contains. 

The COM, SUP, and NEA data files can be used in EQ3NR 
and EQ6 with the Davis equation or the B-dot equation 
because these files contain data specific to a general 
extended Debye-Hiickel model to determine the activity 
coefficients. The Davis and B-dot equations are only valid 
in dilute solutions. 

The COM (composite) file is the largest of the files 
with the information obtained from many sources, including 
those used by the other four data files. The COM file 
represents 886 pure minerals, 852 aqueous species, 76 
gaseous species, 78 chemical elements, and 12 solid 
solutions. It has a temperature range of to 300°C. This 
allows for modeling aqueous solutions with a high degree of 
compositional complexity, such as those of J-13. The COM 
file was produced by Lawrence Livermore National Laboratory. 

The HMW and PIT data files can be used in EQ3NR and EQ6 
when there are solutions containing high concentrations 
(like brine solutions) needing to be analyzed. These data 
use specific semi-empirical equations proposed by Pitzer 






187 
that have been successful in describing the activity- 
coefficients in highly concentrated solutions. ^^ The HMW 
and PIT data files do not contain as many chemical 
components as the others. 

EQ3NR is the speciation-solubility code that is used 
for geochemical modeling. Static calculations based on 
water-chemistry analysis are used to model the thermodynamic 
state of an aqueous solution. The user specifies the 
chemical analysis of the input solution, usually in terms of 
concentrations of dissolved components. These are simply 
the initial concentrations used in the equations. The input 
concentrations are the total concentrations for the species 
and not individual contributions from simple ions, ion 
pairs, and aqueous complexes. However, the data may 
distinguish a dissolved component by oxidation state. For 
instance, Fe""* may have one concentration entered and Fe*** 
another. The pH of the solution is usually entered; 
however, the pHCl may be used. The redox potential (Eh) or 
the oxygen fugacity (pe) may also be used. The redox state 
may alternately be specified by a redox couple. 

A master set of basis species that represent the 
chemical components of the solution are then determined on a 
one-to-one basis from the input concentrations. The strict 
basis is then determined as the minimum number of species 
and includes one species for each element plus one for 
oxidation-reduction and charge balance. An auxiliary basis 



188 
contains species that are related to the strict basis 
through various chemical reactions. Sometimes, a strict 
basis is switched with an auxiliary basis to improve the 
initial starting values for solving the algebraic equations. 

EQ3NR then takes these input concentrations and uses 
them with mass-balance equations and/or charge-balance 
equations. The governing equations are the mass-balance 
equations and charge-balance equations, depending on whether 
concentrations, or redox couples, etc., were entered in the 
input and also depending on the output parameters desired. 
For example, the mass-balance equation is the governing 
equation for an input of total concentration. For the case 
of calcium, the mass-balance equation in terms of molalities 
would be represented by 

%,Ca-2 = ^ca>2 + mca(0H)2(aq)^ "'caCOscaq)'" ^CaHC03+ + ' ' ' (A.l) 

The concentrations of ion-pairs and complexes are 
governed by thermodynamic equilibrium and are represented by 
a mass-action equation for its dissociation. For example, 
the dissociation of calcium carbonate and the corresponding 
mass-action equation would be 

^aC03(3q) ^ Ca^^ + COi^ (A. 2) 

^ ^ _ [3ca'2] [aco^] 

■^CaCOjoq)- r- Y' (^'^) 

l-dcaC03(aq)-l 



.; 189 

where K is the equilibrium constant, and a is the 
thermodynamic activity. The thermodynamic activity is 
related to the molal concentration, m, and the activity 
coefficient, X, by the equation: 

3j = m,- A.,- (A. 4) 

Using the input values and supporting library and data 
files, EQ3NR then uses the standard-state thermodynamic data 
taken from the data file of all the species involved and the 
equations that describe the activity coefficients to compute 
the thermodynamic state of the aqueous solution. For all 
the species, it calculates the concentrations, activity 
coefficients, and thermodynamic activities. 

Because the above equations need to be solved 
simultaneously, EQ3/6 solves equations A.l through A. 4, 
which are intrinsically algebraic, by using a highly 
effective hybrid Newton-Raphson algorithm. The Newton- 
Raphson method is a well known iterative technique for 
solving non-linear systems of algebraic equations. It is 
"perhaps the most widely used of all root-locating 
formulas."" The code uses a first-order algorithm and may 
do some basis switching to optimize the starting estimates 
of the activity coefficients it creates for the Newton- 
Raphson iteration. The activity coefficients of the aqueous 
species are then held constant in each iteration. The 



k w ^ ■••%.'■ - ■■*■ 4, , 



190 
hybrid part comes from readjusting the activity coefficients 
between iterations. 

After going through many iterations, the concentration, 
activity coefficients, and thermodynamic activities of all 
the species are determined. Using these data, it then 
calculates a set of saturation indices and thermodynamic 
affinities of various reactions. The saturation indices are 
measures of the degree of disequilibrium of the reactions: 
the larger the number, the larger the degree of 
disequilibrium. The formula for the saturation indices is 

SI = log I (A. 5) 

where Q is the activity product, and K is the equilibrium 
constant. The thermodynamic affinity of the precipitation 
reaction is represented by the formula 

a = 2.303 RT loq9. = -2.303 RT log:^ (A. 6) 

and the two preceding formulas are related by 

A_ = 2.303 RT SI (A. 7) 

where R is the universal gas constant (8.31441 J K''' mole"''), 
and T is the temperature in Kelvin. The thermodynamic state 
of each aqueous redox couple, the theoretical Eh, pe, oxygen 
fugacity, and redox affinity (Ah) are also calculated when 
there are aqueous redox reactions. EQ3NR next calculates a 



■, ■• 191 

mass balance for each chemical element corresponding to the 
strict-basis species. This mass balance is represented in 
terms of moles. It then computes a charge-balance equation 
based on the charge equivalents. The mass- and charge- 
balance relations obtained are also used by EQ6 to determine 
the number of moles of each strict-basis species as well as 
the oxygen fugacity of the system. If EQ3NR determines a 
charge imbalance, it will pass on to, and be maintained by, 
EQ6. 

The EQ3NR output file is a regurgitation of the input 
plus the output results. The pickup file contains the 
concentrations, affinity coefficients, and thermodynamic 
activities based on mass- and charge-balance equations. The 
pickup file gets added to the bottom of the EQ6 input file, 
and therefore, EQ3NR is "required to initialize the EQ6 
calculations". 2' EQ6 can then be used to model what occurs 
in (1) titration processes (including fluid mixing), 

(2) irreversible reactions in a closed system, 

(3) irreversible reactions in fluid centering flow-through 
systems, and (4) heating and cooling processes. It can also 
solve single-point thermodynamic equilibrium problems. This 
code models both mineral dissolution and growth kinetics. 
The reaction progress or chemical development is driven by 
irreversible reactions, and these reactions are usually the 
dissolution and precipitation of solids and minerals. 



•; " 192 

Modeling one of the four processes mentioned above 
using EQ6 requires supplying the thermodynamic model of an 
aqueous solution, obtained by running EQ3NR, choosing a set 
of irreversible reactions ("reactants") and/or changing the 
temperature/pressures of the system, providing parameters to 
define the rates of these processes, and choosing from among 
the various model options that are available. 

It is important that in EQ3NR, the number of moles of 
solvent water is constant and corresponds to 1 kg. However, 
in EQ6, this is a calculated variable. Therefore, the 
thermodynamic equations in EQ3NR and EQ6 differ, and EQ6 
uses a mass-balance equation for oxygen as the added 
equation for this unknown. The amount of solvent water in 
EQ6 changes, as well as the pH and oxygen fugacity of the 
system; therefore, EQ6 indicates that compounds will 
precipitate when EQ3NR does not. 

Because EQ6 indicates that compounds will precipitate, 
the thermodynamic equations in EQ6 are made in steps, each 
step presuming a phase assemblage. The first is that 
obtained from EQ3NR via the pickup file. EQ6 will then 
check for any phase that is supersaturated. If there are 
supersaturated phases, one of these phases is added to the 
assemblage, and a new calculation is made. A number of new 
unknowns are added, depending on what phase is added. One 
unknown is added if the phase is a pure mineral. The added 
equation for this unknown is the mass-action equation for 



■''■\'.' ; f • « ; 1 .,< y /■'••i 



193 
the pure mineral. However, the number of unknowns will 
increase by the number of end members present in a solid 
solution should that be the phase added. The added 
equations will therefore include a new equation for each end 
member. If the process is successful for the supersaturated 
phases, then it is repeated for all unexcused supersaturated 
phases. 

Similar to EQ3NR, the mathematical expressions for the 
activity coefficients for solid-solution end members are 
used in the mass-action equations that define the solubility 
equilibria. Correction of these activity coefficients is 
done in a Newton-Raphson step. 

The process by which a set of irreversible reactions 
proceeds to thermodynamic equilibrium is represented by the 
reaction path. Chemically speaking, a system is driven as 
the irreversible reactions move toward a state of 
equilibrium. If the aqueous solution is undersaturated with 
respect to a mineral, the thermodynamic driving force is for 
the dissolution of the compound containing that mineral. In 
EQ6, this process is divided into incremental steps of 
adding the mineral. This creates a series of thermodynamic 
equilibrium calculations for systems with increased mass- 
balance totals for each of the elements in the compound 
being dissolved. The reaction path ends when the solution 
is saturated or no more reactant is left. Secondary 



^■■^■<- . - :.-■•.• - 194 

minerals may also form and precipitate during the course of 
this path. 

Reaction-path calculations may take place in "reaction 
progress" mode as relative rates, without reference to time, 
or in "time" mode. Thermodynamic calculations in the 
reaction progress mode are intrinsically algebraic, and as 
such are handled algebraically using iteration variables. 
Choosing time-mode operation is tricky because "the study of 
kinetics of reactions occurring in aqueous geochemical 
systems is still in the pioneering stage. "^^ General 
agreement of the functional form of the rate law does not 
exist. There is also not general agreement in the values of 
the corresponding constants (rate constants, activation 
energies, etc.).^^ Rate-law calculations in the time mode 
are intrinsically differential and as such are handled by 
linear ordinary differential equations. Therefore, if a 
time mode is used, the code must integrate ordinary 
differential equations (the rate laws) . 

EQ6 has three output files that indicate the reaction 
path of a system. These output files indicate the product 
minerals formed, the variations in solution environment 
variables, concentrations of reactants in solution, gases 
produced, and so forth. The three files are the output, 
tab, and pickup files. The output file is usually extremely 
large and cumbersome to work with; however, it gives the 
complete details of what occurred during each step in the 



195 
run. The tab file puts this information in a tabular 
format, which is still quite large and not always convenient 
to work with when analyzing data. The tab file gives all 
the information for all the different phases of the 
different solids, minerals, and gasses. The pickup file is 
capable of being used as the input for another EQ6 run. It 
is convenient to be able to use the pickup file as an input 
file for EQ6 to understand what might occur when the 
temperatures change in the system being analyzed. 



,f. • p il ., '-.^ 



< ^ ''J" ■• • » »J > .^ 



99 



APPENDIX B 
BETA DETECTION SYSTEM 

Technetiuin-99 is a pure beta emitter, which decays to 
'Ru by the emission of a 292 keV B" particle. This process 
is schematically written as 

43TC'' -► 44RU" + B" + iL 

In beta analysis, the beta spectrum is a continuum. It 
appears over the range of zero to the endpoint energy since 
the decay energy is shared between the antineutrino and the 
beta particle, in this case from keV to 292 keV. 

Liquid scintillation analysis (LSA) was chosen for 
determining the ''tc concentration in the samples because 
the sample is counted directly in the liquid scintillator. 
This allows for high efficiency counting of beta particles. 
It also eliminates the problems of self-absorption, 
attenuation by detector windows, and backscattering from the 
detector . 

For LSA, the sample to be counted is dissolved in a 
liquid scintillation cocktail, put in a glass vial, and then 
placed in the counter. All the beta particles emitted from 
the sample must pass through the liquid scintillation 
cocktail where the majority of the beta particles are fully 
stopped. Within the cocktail, the beta particle's energy is 

196 



197 
transferred to the scintillator, and light pulses are 
produced. The light pulses produced enter the 
photomultiplier (PM) tube and bombard the photocathode. The 
photocathode converts the light pulses to low-energy 
electrons, which are accelerated toward the first dynode, 
and many secondary electrons are produced for each incident 
photoelectron. The secondary electrons produced at the 
first dynode are accelerated toward a second similar dynode, 
and more secondary electrons are produced. This process is 
repeated many times, the number of electrons is 
substantially increased, and this amplified electric charge 
is collected at the anode where it serves as a detectable 
electrical signal that is directly proportional to the 
number of original photoelectrons. 

Different liquid scintillation cocktails can greatly 
affect the efficiency of the light pulses produced and hence 
the efficiency of the beta analysis. Therefore, different 
scintillation cocktails were investigated for efficiency; 
Pico Flour LLT was selected for use in this work. Pico 
Flour LLT was chosen because the counting efficiency for 
"Tc/J-13 solutions were between 66-88%, well above the 
other scintillation cocktails. 

With each series of samples counted, a background 
sample was included that contained the scintillation 
cocktail and the same solution as the samples less the '^Tc. 

i .'4 A 

An example of the output obtained from the LSA is shown on 



, 198 

page 199. The plot shows that there were 3222.73 counts per 
minute (CPM) for region B. Region B covers energy ranges 
from 10-250 keV. The first region, Region A, goes from 0-10 
kev. This region is not included for counting the activity 
for "tc because this is the region where the background is 
high, mostly due to electronic noise. The background counts 
obtained from the background sample counted with this sample 
would then be subtracted. 



' ■; V »i- 5 



>>;». >-i -iv i». i . " ^T. ^ "-\ 



<^ ,■ 



'^^if 



i ■. ■ r Ti ^ <„ > ■ ■ •• , f. 



199 



SPECTROGRAPH 2.29 (C) Copyright Packard Instrument Co. I<?a7 09-Api — 93 09:29 

Log VIEW 
5000.0 Y Transform Log max Endpoint 557 keV 

// Active File: TR-6-6A . lOO W NonActive File: PF-TR-6A.1O0 




2000 



Regions of Interest 





LL 


UL 


A: 


0.0 


.10.0 


B: 


10.0 


250. 


C: 


0.0 


2000 


D: 


0.0 


0.0 


E: 


0.0 


0.0 


Fl 


0.0 


O.O 







CPM 




TR-6- 


-6A. 


lOO 


PF-TR-&A 


670 


53 




6.56 


3222 


73 




7.57 


3716 


36 




38.68 


O 


OO 




O . OO 


O 


OO 




o.oo 





OO 




o.oo 






APPENDIX C 
GAMMA DETECTION SYSTEMS 



Technetiuin-95in is a metastable radionuclide with a 
complex decay scheme. Its decay modes include beta-plus, 
electron capture, x-rays, and gamma ray emission. ^^ Unlike 
beta analysis, in gamma analysis each gamma ray produces a 
photopeak at the energy of that gamma ray and, depending on 
the resolution and efficiency of the detector, each peak can 
be discriminated from the others. ■' 

The sodium iodide detector (NaI[Tl])' is an inorganic 
scintillator doped with an activator, thallium.^'' Since a 
gamma ray is an uncharged and highly energetic type of 
radiation, it does not ionize the material it passes 
through. To detect the gamma ray, an interaction must 
occur. There are only three interactions that are important 
relative to gamma spectroscopy: the photoelectric effect, 
compton scattering, and pair production. In each of these 
reactions, the gamma ray transfers energy to electrons. The 
electrons, which have a maximum energy equal to the incident 
gamma, will lose their energy through ionization and 
excitation of atoms within the absorber material. The 



^The theory behind the detectors in this section can be 
found in Reference 57 or any good radiation detection book, 



200 



■ '. . . ' 201 

Nal(Tl) must therefore first convert gamma rays to fast 
electrons before it can detect these electrons. 

When a gamma ray is incident upon the crystal lattice 
of a Nal(Tl), energy is absorbed from its interaction with 
the crystal lattice, and an electron in the crystal's 
valance band is excited into the conduction band. There is 
an activator (Tl) , an impurity in the crystal, that creates 
interstitial states in the normally forbidden band. The 
positive hole left in the normally filled valance band when 
the electron was excited drifts to the location of an 
activator site and ionizes it. The migrating electron in 
the conduction band then drops into this ionized site and 
creates a neutral impurity site with its own excited energy 
state in the forbidden gap. The de-excitation of this 
excited state to the valance band will produce a photon in 
the visible energy range. The site in the forbidden gap is 
necessary because if the electron would de-excite directly 
from the conduction band to the valance band, the energy of 
the resultant photon would be too high to be in the visible 
region; the same holds true for the original gamma ray 
photon. Once the photon in the visible energy range is 
produced, it enters the PM tube, and the same process as 
described in Appendix B occurs. 

The Nal(Tl) detection system was chosen for counting 
the 95"'Tc samples because it has the highest known light 
yield to primary and secondary electrons for gamma analysis, 



■^<- 



202 
and its response is close to linear with respect to the 
incident gamma ray energy. The resolution of the detector 
is not as high as that of germanium detectors. Because the 
'Snirpc was known not to have any impurities, finer resolution 
to discriminate against decay of impurities was not 
necessary. 

The Nal(Tl) was also chosen because of its high atomic 
number of 53, which has a significant influence as to which 
of the three reactions that the gamma ray would undergo to 
produce the electrons. Photoelectric effect is the 
preferred interaction for gamma spectroscopy, and its 
crossection is proportional to z^-^. Technetium-95m has 
multiple gamma rays. The three significant ones are 2 04 keV 
(70%), 838 keV (27%), and 1.04 MeV (4%).^*^ Because the 204 
keV is the most abundant and also falls into the range of 
the photoelectric effect, this is the peak that was used for 
determining the technetium concentration in the analytical 
samples. 

The specific Nal(Tl) detector used for this research 
has the added benefit of being a well-type detector to 
achieve close to a 4 tt geometry, and it is a large crystal, 
4 in. X 5 in., thus having a high efficiency. A sample of 
the output obtained by the analysis of one of the analytical 
samples is shown on page 204. The output shows 348 counts 
per second for the peak used (at energy 204.32). A 
background sample that was counted for the same length of 



203 
time is then subtracted from this. For this sample, the 
count time was only 1 h because the sample was well above 
background. However, for the samples from the sealed vessel 
experiments, 16 h counts were conducted. 






■? ~ T '»'•;■ 



204 



VMS Peak Search Report VI. 9 Generated 15-APR-1994 12:50:05 



Configuration 
Analyses by 
Sample title 
Sample date 
Sample ID 
Sample type 
Detector name , 
Elapsed live time 
Start energy 
Sensitivity 
Critical level 



DKA200: (USER) S4273. CNF; 1 
PEAK V16.4,ENBACK VI . 5 

Tc-95m, F7-J13, and Cu Experiment # MC 
Acquisition date : 



MC-5, S4273 
Liquid Effluent 
S 

01:00:00.00 
20.00 keV 
5.00 
No 



Sample quantity 
Sample geometry 
Detector geometry: 
Elapsed real time : 
End energy 
Gaussian 



15-APR- 
5 . 0000 
Tube 



1994 
nL 



11:48:43 



01:00:46.78 
2560.00 keV 
10.00 



1.3% 



Pk It Energy 



97.84 
204.32 
431.85 
589.44 
813.18 



Area Bkgnd FWHM Channel Left Pw Cts/Sec %Err 



Fit 



28387 

1252004 

6550 

73492 

306707 



213335 14.22 
577893 28.78 
188771 40.95 
215483 43.56 
585265 51.29 



9.78 
20.43 
43.19 
58.94 
81.32 



8 
18 
41 
56 
78 



7.89E+00 
3.48E+02 
1.82E+00 
2.04E+01 



8 8.52E+01 



6 1058.36 147998 230549 55.99 105.84 102 9 4.11E+01 0.6 



Flag: "*" — Peak area was modified by background subtraction 






205 
A high-purity germanium detector (HPGE) is a 
semiconductor detector that is a very high-purity crystal of 
germanium. The general characteristics needed to understand 
the basic means of detecting radiation are explained in the 
discussion that follows. 

In a crystal, there are two bands that electrons can 
occupy: the valance and conduction bands. The valance band 
contains electrons that are bound to specific lattice sites 
within the crystal; in germanium, the electrons are bound by 
a covalent bond. The conduction band contains electrons 
that are not bound and migrate throughout the crystal. 
These two bands are separated by a "forbidden" band gap. A 
crystal contains just enough electrons to fill the valance 
band, and for electrons to enter the conduction band, they 
must gain enough energy (on the order of several eV) to 
cross the gap. Thermal energy can supply a valance electron 
enough energy to excite it into the normally empty 
conduction band, leaving behind an unfilled covalent bond 
site (hole) in the valance band. Under the influence of an 
electric field, the electron migrates in the conduction 
band, and the hole moves in the opposite direction, parallel 
to the electric field applied: this motion of the pair is 
the conductivity of the material. The net drift of 
electrical charges parallel with an electric field is called 
the drift velocity. There is a saturation velocity in which 



' ' _ . 206 

a further increase in electric field does not increase the 
drift velocity. ='• y 

Although very pure, an HPGE contains a few impurities 
in the crystal that tend to be trivalent. These impurities 
are located in regions called acceptor sites. These regions 
are caused by defects in the germanium lattice. Since 
germanium normally has four covalent bonds per atom, a 
trivalent impurity leaves one covalent bond site unfilled 
(and is therefore called a p-type semiconductor) . When an 
electron fills this site, the covalent bond is less than 
that of the others and actually lies in the forbidden gap, 
close to the valance band. Lithium, which is pentavalent, 
is evaporated and diffused onto one side of the germanium 
surface to provide a surface contact that provides an extra 
"donor" electron. This is called the n* surface contact. 

The other side of the germanium crystal must contain a 
p* contact, usually a metal-to-surface barrier junction, and 
is called a noninjecting contact. Where the lithium 
contacts the germanium, an n*-p junction exists. In the n* 
region, there are more conduction electrons than in the p 
region. This electron gradient causes a net diffusion of 
conduction electrons into the less dense p-type material, 
leaving behind immobile positive charges. These conduction 
electrons will then combine with any holes in the p-type 
material. The same argument holds for holes present in the 
p-type material diffusing into the n* region, leaving behind 



207 
an immobile negative charge. An electric field is created 
on either side of the junction because of the space charges, 
and further diffusion is prevented. This standing potential 
in the junction then causes any electrons created in or near 
the junction to be swept back toward the n-type material and 
any holes toward the p-type material. The region is thus 
"depleted" in that the concentration of holes and electrons 
is very small. 

The only significant charges remaining in the depletion 
region are the immobile ionized donor sites and the filled 
acceptor sites. Because these latter charges do not 
contribute to the conductivity, the depletion region 
exhibits a very high resistivity compared with the n- and 
p-type materials on either side of the junction. When a 
positive voltage is applied to the n* contact with respect 
to the p"" surface, the depletion region is widened. It 
begins at the n* edge of the central region and extends 
further into the p region with increasing voltage. This is 
called reverse biasing the crystal. The electric field is 
maximum at the n* side of the p region and decreases to zero 
at the p* side. Any additional increment of voltage applied 
raises the electric field by a constant amount throughout 
the detector. At very high voltage, the depletion region 
extends all the way from the contact to the other side of 
the crystal. This depletion region is the active volume of 
the detector. When radiation creates ionization as it 



208 
passes through this region, electron-hole pairs are created 
that are swept out of the region by the electric field. 
Their motion constitutes a basic electrical signal for 
detection. ' 

The HPGE was chosen because it has the best energy 
resolution for detecting individual radiation peaks. The 
Nal(Tl), although extremely efficient, has poor resolution 
because there are many inefficient steps in converting the 
incident radiation signal into visible light and then 
converting it to an electrical signal. The energy needed to 
produce one photoelectron is approximately 1000 eV, and the 
number of carriers created in a typical radiation 
interaction is usually no more than a few thousand. The 
statistical fluctuation of such a small number places 
limitations on the energy resolution, even under ideal 
circumstances. To reduce this statistical limitation, the 
number of information carriers per pulse needs to be 
increased; the HPGE achieves this. Therefore, a sample 
containing '^""Tc was analyzed using an HPGE well-type 
detector to confirm that '^""Tc was the only radioactive 
constituent present since it could discriminate between very 
close peaks. Also, since '^"Tc is a gamma emitter, its 
precipitation within the column was also traced using an 
HPGE detector with a well-shielded translation column. 






APPENDIX D 
CHEMICAL ASSAY DATA 



CERTIFICATE OF CALIBRATION 
BETA STANDARD SOLUTION 



Radionuclide 
Hair Life: 
Catalog No. ; 
Source ^4o.: 



Tc-99 
(2.13* 0.05) x lO'^S years 
7099 
407-28 



Description of Solution 

a. Mass of solution: 

b. Chemical form: 

c. Carrier content: 

d. Density: 
Radioimpurities 
Radioactive Daughters 

Radionuclide Concentration 



Customer: BATTELLE N.W. LAB. 


P.ONo.: 156053AKL 




Reference Date: August 1 1992 


I2'X>0 PST. 


Contained Radioactivity: 980 


fCi. 


Contained Radioactivity: 36,300 


kB<i 


5.0750 


grams. 


NaTcOt in water 




None added 




0.9982 grnhnL^XfC 


None detected 





None 



193.2 



fiCVpam 



Method of Calibration 

Weighed aliauots of the solution were assayed using a liquid scintillation counter. 



Uncertainty of Measurement 

a. Systematic uncertainty in instrument calibration; + 2. 1% 

b. Random uncertainty in assay: ± 1.3% 

c. Random uncertainty in wcighing(s): + 0.0% 

d. Total uncertainty at the 99% confidence level: + 3.4% 

NIST Traceability 

This calibration is implicitly traceable to the National Institute of Standards and Technolog)-. 
Notes 

1. Nuclear data were taken from "Table of Radioactive Isotopes", edited by Virginia S. Shirley, 1986. 

2. IPL participates in an NIST measurenieni assurance program to establish and maintain implicit 
traceability for a number of nuclides, based on the blind assay(and later NIST certification) of 
Standard Reference Materials (As in NRC Regulatory Guide 4.15). 



SSBD 



QUALITY CONTROL 
Date Signed 



ISOTOPE PRODUCTS LABORATORIES 

1800 No. Keystone Street, 

Burbank, California 91S04 

(8181 843 - 7000 



210 



211 



Po«J-lt"bfandtoxicansmitialitiemo757J *«iiMi«» / 



a! 



y//i/A<> ^pjrzm 



BipT 



B ^M^- 



mi 



^09S76m/ 



e^ 



yfm/ efi/>i</r\j 



flSni 



%/Piy7rPM1/irf)A' 



u%t 



^'^^ ^7J J?£0/? 



M 
o 

L 

I 

a 



0.1 

o 

111 • 



^fElM Products 



QjslomerSMvico 



CATALOG NO. 



MENPKidiiMOnltrf 



Amcum 



^^^^^ I TECHNICAL DATA 

CONC£KTRATION:fl, ZC; mCSinl V&UME: ^.Q/) "* 

RM)«»UCllOCPURrrY: ^ ^f /» 

PfMNOPAL RAOOACTIve CONTAUINANIS: /\)M/C A^TfCTCO 1^11)? 



TOTAiSOLCtS: 



iik 






CAW 



METKOOOFPftEPARATION: ^{Ao/t}2tC) "'*'Tc 
ASSAVDATE: /-/-)-9y ^O^^ ^^^5/'^C 

RADIATION ONSHKLOCO: y<*^ '**» 

CWnw mmiaMWIUUUtUIKM.0UMIOM:1)«|H*(!lill>MMIgii«eMt 
(>a»licM«iM»iVli>l'»n>(<<l'<'riMriMM«>ilaliM>li«iiMk*«rl«m* 
NniuuiitcliK«iatiaa<c«igrti/4><CM(t>m«>iiili*nl>ri<S«klMa»Mr« 
iw > i » yiun» i m mi«iniy»<»<nii i i >t wii. 



UWT€0 WARRANTY: 

Ou rmi NEN PfoteU WKTMb 9* 
notorak «*) tiy k b tx ol mmiuniiiik 
quttytlthilinKolthilxMnl.TtNretsno 
Mtnnty. ti^itss or inpked. « lo Ik 

(wrpgtc The cuilotMr rnsi M<ty Me 
company, ky ntbimd «iA al «iy bnicii 
«l tarrwty. Because ol he Mierefli 
suscepNiitty to ttfawiian ol those 
praduts contMtig tnort taint ndJo- 
aucStto. or «lacli sh«w t nduKuKteaim- 
pautiM iM ol gruin ttun one pe^Mnl 
(IV) pa nMli, notOulion must be laadc 
vidhin IS <)ys ol lectlpl. Fw tt other 
pnxjucis i60^()eriod Ki beAiwed. Al 
cltim ihjl be (Jeemd wweti in the even! 
tw customer Uib ta notily the Minpaiiy 
witw tne gui/aake period The mixiinun 
falilily Vk txeacbof warrwty ilMl tie the 
invoice price lor uU nuleriite awt this 
iatiility shUI te'tadusiire On HM NH 
PiiihmiMlnilfcttebHteiKyMW- 
l«wlW, MiifMtatHy. iMi^coU, or 
ttuunvum- 



iW) 



i 



f 



V^ f .. V 



















212 
















_l ''• 

•^ o 






•^ o 














o. s 






-, 2: 








^, 






^a 


o 




^=^ 


o 












-§^5 


CM 
O 

• 




3^£ 


tN 
O 












9-^ 


o 




r-i "5 


O 












2-< 






W.H 














s° 






S° 






Il 


il 


















(/] 










r<j 




u 

(11 


OJ 




(0 








Silvi 

(%) 


o 




Silv( 
(%) 


O 




JJ 








o 

• 

o 




O 
O 




:al foreign me 
(as oxides) 


in 














osphorus 
(%) 


H 
O 

o 
o 




osphorus 
(%) 


H 
O 
O 

O 




o 

• 

o 






43 






x: 







H 








0^ 




>, 


Oh 






















>1 


a> 




(0 


a) 












(0 


M 




m 


w 












en 


Q) 


H 


tfl 


0) 


H 


>i 


C — 

•H o\o 


CO 




w 


C — 


O 


< 


c -. 


O 


(C 






< 


to o\o 


O 

• 


(0 o\o 


O 




en 




+J 


c 


o 


Q) 


c 


o 


<: 











(0 




T3 


(0 












w 
a, 



u 


s 




5 



^^ 

a 
a 



u 


s 




+j 
















w 

c 

•H 


Lead 
(%) 


o 
o 

• 

o 




ntimony 
& Tin 
(%) 


H 
O 

O 


ntimony 
& Tin 
(%) 


H 
O 

O 


q 


in 




< 






< 






2£ 

M 


o 
o 

• 


















o 






Lead 
(%) 


o 




Lead 
(%) 


in 
o 
















o 

• 

o 




o 
o 




Copper 
(%) 


O 
O 

O 




Iron 
(%) 


in 
o 
o 

• 

o 


Iron 
(%) 


tn 
o 
o 

o 


u 

•H ^^ 




















f\] 














u 






^ 






w a 








Q) 


in 




(u 


o 




^-1 ^ 








as 








as 








< 














u 






u 


CTi 


- 













u 






•H -^ 






•H ^-. 














c g 






c g 














Q) ft 


tn 




0) a 


in 












M a 






cfl a 














^1 ^ 






v^ ^ 














< 






< 






r-' ^ - 






■ ,.,.. ^ " ■■ 




? 












213 





w 






r>j 






^K 






>i1^ 






>-• ja 0) 






ices 
ted 
Ifat 


in 
o 




^"5 3 


• 

o 




Saw 
,5 u •--' 






W Q) 






^ 






a 






>^ 






■p 


0) 




-H rH 


-P 


>1 


-H U 


Q) 


(0 


•H X 


rH 


UJ 


X5 


ft 


I/I 

< 


d c 


g 


nH -H 







o 





Q) 


M 






r-H 






■P 


n 


J3 

u 


(0 
H 


O 
O 

• 


rn 


3 


o 


;3 


W 




o 






c 






c 


U 




(0 


rH 




-p 


Meta 
Pb) 


rH 
O 






• 

o 




Si 






-p 






o 






2H 


o 




o 




H 


o 




O 






-H ^-, 






c e 






<U ft 


CNJ 




W ft 






Li ^ 






< 







C ^ 






■H O 






w 








"""^ 




o\o 




■S Q) o\o 


s..^ 




O 




in 




0^ 


• 

o 


>i 






(C 






(/) 






Ul 






< 


c 




Li 
Q) 


2£ 

H 


en 
0^ 


T3 






5 













0^ 


Q) 




c 


•H ^^ 


<N 





4H o\o 


o 


Li 


rH ^^ 


• 


H 


d 

M 


o 




Nitrogen 

Compounds 

(%) 


o 




o 




• 

o 



REFERENCES 

1. J. Vida, The Chemical Behavior of Technetium During the 
Treatment of High-Level Radioactive Waste. Institute 
for Hot Chemistry, Karlsruhe Nuclear Research Center, 
KFK 4642, Translated from German by JR Jewett, 
Westinghouse Hanford Company, Richland, Washington PNL- 
TR-497 (1994) . 

2. J. A. Rard, Critical Review of the Chemistry and 
Thermodynamics of Technetium and Some of Its Inorganic 
Compounds and Aqueous Species, UCRL 5344 0, Lawrence 
Livermore National Laboratory, Livermore, California 
(1983) . 

3. F.W. Walker, J.R. Parrington, and F. Feiner, Nuclides 
and Isotopes, Fourteenth Edition, General Electric 
Company, Nuclear Energy Operations, San Jose, 
California (1989) . 

4. P. Cohen, Water Coolant Technology of Power Reactors, 
Westinghouse Electric Corporation, ISBN: 0-89448-020-0 
American Nuclear Society (1980) . ' 

5. J. Weisman, Elements of Nuclear Reactor Design, Robert 
E. Krieger Publishing Company, Malabar, Florida (1983). 

6. U.S. DOE The Nuclear Waste Policy Act of 1982, 42 
U.S.C. 10101 (This act consists of the Act of' January 
7, 1983 (Public Law 97-425 Stat. 2201), as amended by' 
P.L. 100-203 (December 22, 1987) and P.L. 102-486 (The 
Energy Act of 1992, October 24, 1992). 

7. C.N. Wilson and W.J. Gray, "Measurement of Soluble 
Nuclide Dissolution Rates from Spent Fuel", Scientific 
Basis for Nuclear Waste Management XIV, Virginia M 
Oversby and Paul W. Brown, Editors., Materials Research 
Symposium Proceedings, Materials Research Society, 
Pittsburgh, Pennsylvania, Volume 176, pp. 489-498' 
(1990) . ' ff c 

8. C.N. Wilson, Results from NNWSI Series 3 Spent Fuel 
Dissolution Tests, PNL-7170, Pacific Northwest 
Laboratory, Richland, Washington (1990). 



214 



215 

9. M. Benedict, T.H. Pigford, and H.W. Levi, Nuclear 
Chemical Engineering, Second Edition, ISBN 0-07-004531- 
3 McGraw-Hill Book Company, New York (1981) 

10. C.N. Wilson, Results from Cycles 1 and 2 NNWSI Series 2 
Spent Fuel Dissolution Tests, HEDL-TME 85-22 UC-70, 
Hanford Engineering Development Laboratory, Richland, 
Washington (1987). 

11. C.N. Wilson, "Summary of Results from the Series 2 and 
Series 3 NNWSI Bare Fuel Dissolution Tests," Scientific 
Basis for Nuclear Waste Management XI, M.J. Apted and 
R.E. Westerman, Editors., Materials Research Symposium 
Proceedings, Materials Research Society, Pittsburgh, 
PA, Volume 112, pp. 473-483 (1987). 

12. C.N. Wilson and H.F. Shaw, "Experimental Study of the 
Dissolution of Spent Fuel at 85°C in Natural Ground 
Water," Scientific Basis for Nuclear Waste Management 
X, J.K. Bates and W.B. Seefeldt, Editors., Materials 
Research Symposium Proceedings 84, Materials Research 
Society, Pittsburgh, Pennsylvania, pp. 123-130 (1986) . 

13. C.N. Wilson and C.J. Bruton, "Studies on Spent Fuel 
Dissolution Behavior Under Yucca Mountain Repository 
Conditions," Nuclear Waste Management III, G.B. 
Mellinger, Ed,, Ceramic Transactions, American Ceramic 
Society, Westerville, Ohio, Volume 9, pp. 423-441 
(1990). 

14. V.M. Oversby and H.F. Shaw, Spent Fuel Performance 
Data: An Analysis of Data Relevant to the NNWSI 
Project, UCID 2 092 6, Lawrence Livermore National 
Laboratory, Livermore, California (1987) . 

15. Office of Civilian Radioactive Waste Management, 
Nuclear Waste Policy Act (Section 113), Site 
Characterization Plan, Yucca Mountain Site, Nevada 
Research and Development Area, Nevada, DOE/RW-0199, 
U.S. Department of Energy (1988). 

16. V.M. Oversby, Important Radionuclides in High Level 
Waste Disposal : Determination Using a Comparison of the 
EPA and NRC Regulations , UCRL-94222, Lawrence Livermore 
National Laboratory, Livermore, California (1986) . 

17. P.w. Atkins, Physical Chemistry, Third Edition, ISBN 
Number 0-7167-1749-2, W.H. Freeman and Company, New 
York (1986). 



■ ■ --^ 216 

18. D.C. Harris, Quantitative Chemical Analysis, Second 
Edition, ISBN Number 0-716-71817-0, W.H. Freeman and 
Company, New York (1987) . 

19. P.H. Sisler, R.D. Dresdner, and W.T. Mooney, Chemistry, 
A SYSTEMATIC APPROACH, ISBN Number 0-19-502630-6, 
Oxford University Press, Oxford, New York (1980) . 

20. R.E. Meyer, W.D. Arnold and F.I. Case, Valance Effects 
on Solubility and Sorption: The Solubility of Tc(IV) 
Oxides, NUREG/CR-4309 ORNL 6199, Oak Ridge National 
Laboratory, Martin Marietta Energy Systems, Oak Ridge, 
Tennessee (1985) . 

21. R.E. Meyer, W.D. Arnold and F.I. Case, Valence Effects 
on the Sorption of Nuclides on Rocks and Minerals, 
NUREG/CR-3389/ORNL 5978, Oak Ridge National Laboratory, 
Martin Marietta Energy Systems, Oak Ridge, Tennessee 
(1983) . 

22. V. Statler, University of Florida, P. Cloke, Science 
Application International Corporation, and J. Tulenko 
University of Florida, "Determination of a Method to 
Inhibit Technetium Migration at the Proposed Federal 
High Level Waste Repository at Yucca Mountain, Nevada," 
1990 American Nuclear Society Annual Winter Meeting, 
Nashville, Tennessee (1990). 

23. V. Statler, University of Florida, Investigation of the 
Effects of Iron, Tin, Stannous Chloride, Copper, 
Manganese and Vanadium on the Retardation of Technetium 
in J-13 Well Water, Master's Thesis (1989). 

24. V. Statler and W, Ellis, University of Florida, 
"Determination of a Method to Inhibit Technetium 
Migration at a Geologic Repository," Third 
International High Level Waste Management Symposium, 
Las Vegas, Nevada (1992). » 

25. T.J. Wolery, EQ3NR: A Computer Program for Geochemical 
Aqueous Speciation-Solubility Calculations: User's 

. Guide and Documentation, UCRL 53414, Lawrence Livermore 
National Laboratory, Livermore, California (1983) . 

26. T.J. Wolery, EQ3NR, A Computer Program for Geochemical 
Aqueous Speciation-Solubility Calculations: Theoretical 
Manual, User's Guide, and Related Documentation 
(Version 7.0), UCRL-MA-1 10662 PT III, Lawrence 
Livermore National Laboratory, Livermore, California 
(1992) . 



.• ; : ■:,, ' 2i7 

27. T.J. Wolery and S.A. Daveler, EQ6, A Computer Program 
for Reaction Path Modeling of Aqueous Geochemical 
Systems: Theoretical Manual, User's Guide, and Related 
Documentation (Version 7.0), UCRL-MA-110662 PT IV, 
Lawrence Livermore National Laboratory, Livermore, 
California (1992) . 

28. J.M. Delany, I Puigdomenech and T.J. Wolery, 
Precipitation Kinetics Option for the EQ6 Geochemical 
Reaction Path Code, UCRL 53 64 2, Lawrence Livermore 
National Laboratory, Livermore, California (1986) . 

29. T.J. Wolery, EQ3/6, A Software Package for Geochemical 
Modeling of Aqueous Systems: Package Overview and 
Installation Guide (Version 7.0), UCRL-MA-110662 PT I, 
Lawrence Livermore National Laboratory, Livermore, 
California (1992) . 



30 



35, 



B.P. McGrail and W.J. Gray, conversation. Pacific 
Northwest National Laboratory, Richland, Washington 
(2/95). 



31. FT-4C Operators Manual, Test Materials, Certifications, 
Cortest Systems, Inc., Mountain View, California, 
(1987). 

32. W.P. Freeborn and W.B. White, "The role of Eh in 
Nuclear Waste Form Dissolution," Scientific Basis for 
Nuclear Waste Management VII, G.L. McVay, Ed., 
Materials Research Symposium Proceedings Volume 26, 
Elsevier Science Publishing Company, Inc., New York, 
New York, pp. 719-726 (1983). 

33. CM. Jantzen, "Methods of Simulating Low Redox 
Potential (Eh) for a Basalt Repository," Scientific 
Basis for Nuclear Waste Management VII, G.L. McVay, 
Ed., Materials Research Symposium Proceedings Volume 
26, Elsevier Science Publishing Company, Inc., New 
York, New York, pp. 613-621 (1983). 

34. Handbook of Chemistry and Physics, Thirtieth Edition, 
Chemical Rubber Publishing Company (1947). 

R.M. Guppy and A. Atkinson, Experimental Investigation 
of Factors Affecting the Control of Redox Conditions 
Withxn a Radwaste Repository , DOE/HMIP/RR/91/043 , 
Department of the Environment: HMIP-Commissioned' 
Research, Harwell Laboratory, Didcot, Oxfordshire, U.K. 
(1991) . 



218 

36. B. Gu and R.K. Schulz, Anion Retention in Soil: 
Possible Application to Reduce Migration of Buried 
Technetium and Iodine, NUREG/CR 54 64 Department of Soil 
Science, University of California, Berkeley, California 
(1991) . 

37. Dr. Dhanpat Rai, Scientist, Biochemistry Resources 
department at Pacific Northwest National Laboratory, 
Richland, Washington. Conversation, January 1996. 

38. D. Rai and J.L. Ryan, "Neptunium (IV) Hydrous Oxide 
Solubility under Reducing and Carbonate Conditions," 
Inorganic Chemistry, 1985, American Chemical Society, 
Volume 24, pp. 247-251 (1985). 

39. Dr. Walter Gray, Pacific Northwest National Laboratory, 
Richland, Washington. Conversation, January 1996. 

40. L.J. Korb, D.L. Olson, Co-chairmen, Metals Handbook® 
Ninth Edition, Volume 13, Corrosion, ASM International, 
Metals Park, Ohio (1987) . 

41. D. I Sherwood, Application of the Ruthenium and 
Technetium Thermodynamic Data Bases Used in the EQ3/6 
Geochemical Codes, UCRL 53594, Lawrence Livermore 
National Laboratory, Livermore, California (1985) 

42. G.H. Cartledge, "The Mechanisms of the Inhabitation of 
Corrosion by the Pertechnetate Ion. I. The Origin and 
Nature of Reaction Products", Journal of Physics and 
Chemistry, Volume 59, pp 979-984 (1955) 

43. G.H. Cartledge, "Free Energies of Formation of Hydrous 
Oxides of Technetium in Its Lower Valencies", in 
Journal of Electrochemical Society, Volume 118 Number 
2, pages 231-236 (1971) 

44. G.H. Cartledge, "The Electrochemical Behavior of 
Technetium and Iron Containing Technetium", in Journal 
of Electrochemical Society, Volume 118 Number 11. oaaes 
1752-1758 (1971) ' t^ ^ 

45. R.E. Meyer, W.D. Arnold, F.I. Case and G.D. O'Kelley, 
Thermodynamic Properties of Tc(IV) Oxides: Solubilities 
and the Electrode Potential of the Tc (VII) /Tc (IV) Oxide 
Couple, NUREG/CR 5108 ORNL 6480, Oak Ridge National 
Laboratory, Oak Ridge, Tennessee (1988) 



,,, 219 

46. R.E. Meyer, W.D. Arnold, F.I. Case and G.D. O'Kelley, 
Thermodynamics of Technetium Related to Nuclear Waste 
Disposal. Solubilities of Tc(IV) and the Electrode 
Potential of Tc (VII) /Tc (IV) Oxide Couple, NUREG/CR 5235 
ORNL 6503, Oak Ridge National Laboratory, Oak Ridge, 
Tennessee (1989) 

47. R.C. Weast, CRC Handbook of Chemistry and Physics, 
Sixty-fourth Edition, CRC Press Incorporated, Boca 
Raton, Florida (1984). 

48. N.N. Greenwood and A. Earnshaw, Chemistry of the 
Elements, Department of Inorganic and Structural 
Chemistry, University of Leeds, U.K., Pergamon Press 
(1984) 

49. CM. Jantzen, "Effects of Eh (oxidation potential) on 
borosilicate waste glass durability," in Advances in 
Ceramics, Nuclear Waste Management, G.G. Wicks and 
W.A. Ross, Editors, American Ceramic Society, Inc., 
Columbus, Ohio, Volume 8 (1984) . 

50. R.C. Weast, CRC Handbook of Chemistry and Physics, 
Sixty-fourth Edition, CRC Press Incorporated, Boca 
Raton, Florida (1984) . 

51. C.N. Wilson and C.J. Bruton, "Studies on Spent Fuel 
Dissolution Behavior Under Yucca Mountain Repository 
Conditions", Nuclear Waste Management III, G.B. 
Mellinger, Ed., Ceramic Transactions, American Ceramic 
Society, Westerville, Ohio, Volume 9, pp. 423-441 
(1990) 

52. Chart of the Nuclides, General Electric Company, 
Nuclear Energy Operations, San Jose, California, 
(1984). 

53. M.D. Williams, V.R. Vermeul, M. Oostrom, J.C. Evans, 
J.S Fruchter, J.D. Istok, M.D. Humphrey, D.C. Lanigan, 
J.E. Szecsody, M.D. White, T.W. Wistma, and C.R. Cole, 
Anoxic Plume Attenuation in a Fluctuating Water Table 
System: Impact of 100-D Area In Situ Redox Manipulation 
on Downgradient Dissolved Oxygen Concentrations , PNNL- 
12192, UC-2000, Pacific Northwest National Laboratory, 
Richland, Washington (1999) 

54. Technetium Attenuation in the Vadose Zone: Role of 
Mineral Interactions , Investigator N. J. Hess, 
Collaborator S.D. Conradson, LALP-99-194, Los Alamos 
National Laboratory, Los Alamos, New Mexico 



220 



55. Steven C. Chapra and Raymond P. Canale, Numerical 
Methods for Engineers with Personal Computer 
Applications , ISBN 0-07-010664-9 McGraw Hill Book 
Company, New York (1985). 

56. Radiological Health Handbook, Compiled and edited by 
the bureau of Radiological Health, U.S. Department of 
Health, Education, and Welfare, Public Health Service, 
Food and Drug Administration, Bureau of Radiological 
Health, Rockville, Maryland, Revised Edition (1970) 

57. G.F. Knoll, Radiation Detection and Measurement, ISBN 
Number 0-471-49545-X, John Wiley and Sons, New York, 
New York, (1979) 









BIOGRAPHICAL SKETCH 

VirLynda DeMartino was born in New York, New York. She 
knew by the age of nine that she wanted to be a nuclear 
engineer and, therefore, went to Saunders Trades and 
Technical High School where she obtained a Regents High 
School Diploma and graduated with honors in Chemical 
Technology. She obtained her Bachelor of Science in 
Engineering degree from the University of Florida's 
Department of Nuclear Engineering Sciences in May 1988, and 
in December 1989 she obtained her Master of Engineering 
degree from the same department. 

Mrs. DeMartino has worked in the Neutron Activation 
Analysis laboratory, became Qualified Second Person Reactor 
Operator for the University of Florida Training Reactor, and 
interned at Science Application International Corporation in 
Las Vegas, Nevada, and Pacific Northwest National Laboratory 
in Richland, Washington. She is experienced in the 
following: general and radioanalytical laboratory conduct of 
operations, R&D of radioactive waste management 
investigations, assembly and use of radiation detection 
systems and flow-through experimentation systems, DOS and 
Macintosh operating systems and programs, and Windows 
programs. She is familiar with Fortran programming, and 

221 



■ : : - 222 

0S2, UNIX, and VAX operating systems. She has authored/co- 
authored five publications under the surnames Statler and 
DeMartino. • " 

She has received the following Honors: Associated 
Western Universities, Inc., Northwest Division Fellowship; 
Northwest College and University Association for Science 
Fellowship; U.S. Department of Energy High Level Waste 
Management Fellowship; Innovative Nuclear Power Operations 
Scholarship; American Nuclear Society (ANS) State Section 
Scholarship (2); Who's Who Among Students in American 
Universities and Colleges (2); University of Florida's 
President Award (2); College of Engineering Outstanding 
Leadership Award; College of Engineering Outstanding Service 
Award; ANS National Outstanding Service Award; ANS Student 
Branch Outstanding Service Award; Benton Engineering Council 
(BEC) Outstanding Officer Award; Outstanding Service to the 
Engineers' Fair (2); and the Columbia Scholastic Medal for 
Outstanding Service. She was on the University of Florida's 
Dean's List. She also has been a member of the following 
societies: ANS, BEC, Epsilon Lambda Chi, Florida Engineering 
Society, Omicron Delta Kappa, and the Society of Women 
Engineers. 

Mrs. DeMartino has held the following University of 
Florida extracurricular activity positions: 1990 Engineers' 
Fair Chairman; 1989 Engineers' Fair Design Contest 
Coordinator; BEC Programs Director (in charge of the 



. t ^■■■■■ 

223 



College of Engineering's Homecoming Float, Holiday Cheer 
from a Gator Engineer, the College of Engineering's Pre- 
Engineering Colloquium, and High School Outreach) ; BEC 
Interim President; ANS 1991 Eastern Regional Student 
Conference (ERSC) Co-Chairman; ANS Student Branch President; 
ANS Student Branch Public Relations Chairman; ANS 1985 ERSC 
Housing Coordinator; contributed many papers at ANS related 
meetings; Epsilon Lambda Chi fraternity secretary; and 
Omicron Delta Kappa fraternity tapping committee. 









,. ■ ■ , ■■■■■-> . *'j «".,»"" ' 






I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is fully 
adequate, in scope and quality, as a dissertation for the degree of 
Doctor of Philosophy. 




Samim Anghaie, cjtiairman 
Professor of Nuclear and 
Radiological Engineering 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is fully 
adequate, in scope and quality, as a dissertation for the degree of 
Doctor of Philosophy. 




Walter J. tcray, CocJS'airman 
Senior Scientist, Pacific 
Northwest National Laboratory 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is fully 
adequate, in scope and quality, as a dissertation for the degree of 
Doctor of Philosophy. 




LcAnI:y^ 



flenko 
)fessor of Nuclear 
ind Radiological Engineering 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is fully 
adequate, in scope and quality, as a dissertation for the degree of 
Doctor of Philosophy. \ 




loessow 
isor Emeritus of Nuclear 
Radiological Engineering 



I certify that I have read this study and that in my opinion it 
conforms to acceptable standards of scholarly presentation and is fully 
adequate, in scope and quality, as a dissertation for the degree of 
Doctor of Philosophy. 




(XM/yJ^ ' \ 



Hanrahan 
ssor of Chemistry 



'"% .'*'; 



This dissertation was submitted to the Graduate Faculty of the 
College of Engineering and to the Graduate School and was accepted as 
partial fulfillment of the requirements for the degree of Doctor of 
Philosophy. 



December 2000 



A 



^.. 



Jack Ohanian 

Dean, College of Engineering 



Winfred M. Philips 
Dean, Graduate School 



■:'-.3fc 



i.;-* 



f^ 



*■■--,. V, 






y"i-'. 



..?•,.»*: 




.0^1151 



UNIVERSITY OF FLORIDA 



3 1262 08555 1686